diff --git a/internal/admin/dashboard/templates/x-icon.html b/internal/admin/dashboard/templates/x-icon.html
new file mode 100644
index 000000000..db1fd85b5
--- /dev/null
+++ b/internal/admin/dashboard/templates/x-icon.html
@@ -0,0 +1,6 @@
+{{define "x-icon"}}
+
+{{end}}
diff --git a/internal/admin/handler.go b/internal/admin/handler.go
index a6e103dab..fc6e16dce 100644
--- a/internal/admin/handler.go
+++ b/internal/admin/handler.go
@@ -377,9 +377,23 @@ func (h *Handler) DailyUsage(c *echo.Context) error {
// @Failure 401 {object} core.GatewayError
// @Router /admin/api/v1/usage/models [get]
func (h *Handler) UsageByModel(c *echo.Context) error {
- return usageSliceResponse(c, h.usageReader, func(ctx context.Context, params usage.UsageQueryParams) ([]usage.ModelUsage, error) {
- return h.usageReader.GetUsageByModel(ctx, params)
- })
+ if h.usageReader == nil {
+ return c.JSON(http.StatusOK, []usage.ModelUsage{})
+ }
+
+ params, err := parseUsageParams(c)
+ if err != nil {
+ return handleError(c, err)
+ }
+
+ values, err := h.usageReader.GetUsageByModel(c.Request().Context(), params)
+ if err != nil {
+ return handleError(c, err)
+ }
+ if values == nil {
+ values = []usage.ModelUsage{}
+ }
+ return c.JSON(http.StatusOK, values)
}
// UsageLog handles GET /admin/api/v1/usage/log
@@ -392,7 +406,7 @@ func (h *Handler) UsageByModel(c *echo.Context) error {
// @Param start_date query string false "Start date (YYYY-MM-DD)"
// @Param end_date query string false "End date (YYYY-MM-DD)"
// @Param model query string false "Filter by model name"
-// @Param provider query string false "Filter by provider"
+// @Param provider query string false "Filter by provider name or provider type"
// @Param user_path query string false "Filter by tracked user path subtree"
// @Param cache_mode query string false "Cache mode filter: uncached, cached, all (default uncached)"
// @Param search query string false "Search across model, provider, request_id, provider_id"
@@ -500,15 +514,15 @@ func (h *Handler) CacheOverview(c *echo.Context) error {
// @Param days query int false "Number of days (default 30)"
// @Param start_date query string false "Start date (YYYY-MM-DD)"
// @Param end_date query string false "End date (YYYY-MM-DD)"
-// @Param model query string false "Filter by model name"
-// @Param provider query string false "Filter by provider"
+// @Param requested_model query string false "Filter by requested model selector"
+// @Param provider query string false "Filter by provider name or provider type"
// @Param method query string false "Filter by HTTP method"
// @Param path query string false "Filter by request path"
// @Param user_path query string false "Filter by tracked user path subtree"
// @Param error_type query string false "Filter by error type"
// @Param status_code query int false "Filter by status code"
// @Param stream query bool false "Filter by stream mode (true/false)"
-// @Param search query string false "Search across request_id/model/provider/method/path/error_type"
+// @Param search query string false "Search across request_id/requested_model/provider/method/path/error_type"
// @Param limit query int false "Page size (default 25, max 100)"
// @Param offset query int false "Offset for pagination"
// @Success 200 {object} auditlog.LogListResult
@@ -531,18 +545,23 @@ func (h *Handler) AuditLog(c *echo.Context) error {
return handleError(c, err)
}
+ requestedModel := c.QueryParam("requested_model")
+ if requestedModel == "" {
+ requestedModel = c.QueryParam("model")
+ }
+
params := auditlog.LogQueryParams{
QueryParams: auditlog.QueryParams{
StartDate: dateRange.StartDate,
EndDate: dateRange.EndDate,
},
- Model: c.QueryParam("model"),
- Provider: c.QueryParam("provider"),
- Method: strings.ToUpper(c.QueryParam("method")),
- Path: c.QueryParam("path"),
- UserPath: userPath,
- ErrorType: c.QueryParam("error_type"),
- Search: c.QueryParam("search"),
+ RequestedModel: requestedModel,
+ Provider: c.QueryParam("provider"),
+ Method: strings.ToUpper(c.QueryParam("method")),
+ Path: c.QueryParam("path"),
+ UserPath: userPath,
+ ErrorType: c.QueryParam("error_type"),
+ Search: c.QueryParam("search"),
}
if sc := c.QueryParam("status_code"); sc != "" {
diff --git a/internal/admin/handler_test.go b/internal/admin/handler_test.go
index 63909ac92..7316da3b1 100644
--- a/internal/admin/handler_test.go
+++ b/internal/admin/handler_test.go
@@ -441,6 +441,37 @@ func TestUsageByModel_Success(t *testing.T) {
}
}
+func TestUsageByModel_PreservesProviderName(t *testing.T) {
+ reader := &mockUsageReader{
+ modelUsage: []usage.ModelUsage{
+ {Model: "gpt-4o", Provider: "openai", ProviderName: "primary-openai", InputTokens: 100, OutputTokens: 25},
+ },
+ }
+ h := NewHandler(reader, nil)
+ c, rec := newHandlerContext("/admin/api/v1/usage/models?days=30")
+
+ if err := h.UsageByModel(c); err != nil {
+ t.Fatalf("unexpected error: %v", err)
+ }
+ if rec.Code != http.StatusOK {
+ t.Fatalf("expected 200, got %d", rec.Code)
+ }
+
+ var models []usage.ModelUsage
+ if err := json.Unmarshal(rec.Body.Bytes(), &models); err != nil {
+ t.Fatalf("failed to unmarshal: %v", err)
+ }
+ if len(models) != 1 {
+ t.Fatalf("expected 1 entry, got %d", len(models))
+ }
+ if models[0].ProviderName != "primary-openai" {
+ t.Fatalf("ProviderName = %q, want %q", models[0].ProviderName, "primary-openai")
+ }
+ if models[0].Provider != "openai" {
+ t.Fatalf("Provider = %q, want %q", models[0].Provider, "openai")
+ }
+}
+
func TestUsageByModel_Error(t *testing.T) {
reader := &mockUsageReader{
modelUsageErr: errors.New("db failure"),
@@ -525,6 +556,51 @@ func TestUsageLog_Success(t *testing.T) {
}
}
+func TestUsageLog_PreservesProviderName(t *testing.T) {
+ now := time.Now().UTC()
+ reader := &mockUsageReader{
+ usageLog: &usage.UsageLogResult{
+ Entries: []usage.UsageLogEntry{
+ {
+ ID: "1",
+ RequestID: "req-1",
+ Model: "gpt-4o",
+ Provider: "openai",
+ ProviderName: "primary-openai",
+ Timestamp: now,
+ InputTokens: 100,
+ TotalTokens: 100,
+ },
+ },
+ Total: 1,
+ Limit: 50,
+ },
+ }
+ h := NewHandler(reader, nil)
+ c, rec := newHandlerContext("/admin/api/v1/usage/log?days=30")
+
+ if err := h.UsageLog(c); err != nil {
+ t.Fatalf("unexpected error: %v", err)
+ }
+ if rec.Code != http.StatusOK {
+ t.Fatalf("expected 200, got %d", rec.Code)
+ }
+
+ var result usage.UsageLogResult
+ if err := json.Unmarshal(rec.Body.Bytes(), &result); err != nil {
+ t.Fatalf("failed to unmarshal: %v", err)
+ }
+ if len(result.Entries) != 1 {
+ t.Fatalf("expected 1 entry, got %d", len(result.Entries))
+ }
+ if result.Entries[0].ProviderName != "primary-openai" {
+ t.Fatalf("ProviderName = %q, want %q", result.Entries[0].ProviderName, "primary-openai")
+ }
+ if result.Entries[0].Provider != "openai" {
+ t.Fatalf("Provider = %q, want %q", result.Entries[0].Provider, "openai")
+ }
+}
+
func TestUsageLog_Error(t *testing.T) {
reader := &mockUsageReader{
usageLogErr: core.NewProviderError("test", http.StatusBadGateway, "upstream failed", nil),
@@ -591,15 +667,15 @@ func TestAuditLog_Success(t *testing.T) {
logResult: &auditlog.LogListResult{
Entries: []auditlog.LogEntry{
{
- ID: "log-1",
- Timestamp: now,
- DurationNs: 12_000_000,
- Model: "gpt-4o",
- Provider: "openai",
- StatusCode: 200,
- RequestID: "req-1",
- Method: http.MethodPost,
- Path: "/v1/chat/completions",
+ ID: "log-1",
+ Timestamp: now,
+ DurationNs: 12_000_000,
+ RequestedModel: "gpt-4o",
+ Provider: "openai",
+ StatusCode: 200,
+ RequestID: "req-1",
+ Method: http.MethodPost,
+ Path: "/v1/chat/completions",
Data: &auditlog.LogData{
RequestBody: map[string]any{
"model": "gpt-4o",
@@ -644,6 +720,50 @@ func TestAuditLog_Success(t *testing.T) {
}
}
+func TestAuditLog_PreservesProviderName(t *testing.T) {
+ now := time.Now().UTC()
+ reader := &mockAuditReader{
+ logResult: &auditlog.LogListResult{
+ Entries: []auditlog.LogEntry{
+ {
+ ID: "log-1",
+ Timestamp: now,
+ RequestedModel: "smart",
+ ResolvedModel: "primary-openai/gpt-4o",
+ Provider: "openai",
+ ProviderName: "primary-openai",
+ StatusCode: 200,
+ },
+ },
+ Total: 1,
+ Limit: 25,
+ },
+ }
+ h := NewHandler(nil, nil, WithAuditReader(reader))
+ c, rec := newHandlerContext("/admin/api/v1/audit/log?days=7")
+
+ if err := h.AuditLog(c); err != nil {
+ t.Fatalf("unexpected error: %v", err)
+ }
+ if rec.Code != http.StatusOK {
+ t.Fatalf("expected 200, got %d", rec.Code)
+ }
+
+ var result auditlog.LogListResult
+ if err := json.Unmarshal(rec.Body.Bytes(), &result); err != nil {
+ t.Fatalf("failed to unmarshal: %v", err)
+ }
+ if len(result.Entries) != 1 {
+ t.Fatalf("expected 1 entry, got %d", len(result.Entries))
+ }
+ if result.Entries[0].ProviderName != "primary-openai" {
+ t.Fatalf("ProviderName = %q, want %q", result.Entries[0].ProviderName, "primary-openai")
+ }
+ if result.Entries[0].Provider != "openai" {
+ t.Fatalf("Provider = %q, want %q", result.Entries[0].Provider, "openai")
+ }
+}
+
func TestAuditLog_WithFilters(t *testing.T) {
reader := &mockAuditReader{
logResult: &auditlog.LogListResult{
@@ -664,8 +784,8 @@ func TestAuditLog_WithFilters(t *testing.T) {
t.Errorf("expected 200, got %d", rec.Code)
}
- if reader.lastQuery.Model != "gpt-4" {
- t.Errorf("expected model filter gpt-4, got %q", reader.lastQuery.Model)
+ if reader.lastQuery.RequestedModel != "gpt-4" {
+ t.Errorf("expected requested model filter gpt-4, got %q", reader.lastQuery.RequestedModel)
}
if reader.lastQuery.Provider != "openai" {
t.Errorf("expected provider filter openai, got %q", reader.lastQuery.Provider)
@@ -795,6 +915,42 @@ func TestAuditConversation_Success(t *testing.T) {
}
}
+func TestAuditConversation_PreservesProviderName(t *testing.T) {
+ now := time.Now().UTC()
+ reader := &mockAuditReader{
+ conversationResult: &auditlog.ConversationResult{
+ AnchorID: "log-2",
+ Entries: []auditlog.LogEntry{
+ {ID: "log-1", Timestamp: now.Add(-time.Minute), ResolvedModel: "primary-openai/gpt-4o", Provider: "openai", ProviderName: "primary-openai", Path: "/v1/responses"},
+ {ID: "log-2", Timestamp: now, ResolvedModel: "primary-openai/gpt-4o", Provider: "openai", ProviderName: "primary-openai", Path: "/v1/responses"},
+ },
+ },
+ }
+ h := NewHandler(nil, nil, WithAuditReader(reader))
+ c, rec := newHandlerContext("/admin/api/v1/audit/conversation?log_id=log-2")
+
+ if err := h.AuditConversation(c); err != nil {
+ t.Fatalf("unexpected error: %v", err)
+ }
+ if rec.Code != http.StatusOK {
+ t.Fatalf("expected 200, got %d", rec.Code)
+ }
+
+ var result auditlog.ConversationResult
+ if err := json.Unmarshal(rec.Body.Bytes(), &result); err != nil {
+ t.Fatalf("failed to unmarshal: %v", err)
+ }
+ if len(result.Entries) != 2 {
+ t.Fatalf("expected 2 entries, got %d", len(result.Entries))
+ }
+ if result.Entries[0].ProviderName != "primary-openai" {
+ t.Fatalf("ProviderName = %q, want %q", result.Entries[0].ProviderName, "primary-openai")
+ }
+ if result.Entries[0].Provider != "openai" {
+ t.Fatalf("Provider = %q, want %q", result.Entries[0].Provider, "openai")
+ }
+}
+
func TestAuditConversation_MissingLogID(t *testing.T) {
reader := &mockAuditReader{}
h := NewHandler(nil, nil, WithAuditReader(reader))
diff --git a/internal/aliases/provider.go b/internal/aliases/provider.go
index dc290b83d..fce30274f 100644
--- a/internal/aliases/provider.go
+++ b/internal/aliases/provider.go
@@ -156,6 +156,13 @@ func (p *Provider) GetProviderType(model string) string {
return p.inner.GetProviderType(model)
}
+func (p *Provider) GetProviderName(model string) string {
+ if named, ok := p.inner.(core.ProviderNameResolver); ok {
+ return strings.TrimSpace(named.GetProviderName(model))
+ }
+ return ""
+}
+
func (p *Provider) ModelCount() int {
if counted, ok := p.inner.(interface{ ModelCount() int }); ok {
return counted.ModelCount()
diff --git a/internal/auditlog/auditlog.go b/internal/auditlog/auditlog.go
index 7c442ce05..b352a7825 100644
--- a/internal/auditlog/auditlog.go
+++ b/internal/auditlog/auditlog.go
@@ -47,9 +47,10 @@ type LogEntry struct {
DurationNs int64 `json:"duration_ns" bson:"duration_ns"`
// Core fields (indexed for queries)
- Model string `json:"model" bson:"model"`
+ RequestedModel string `json:"requested_model" bson:"requested_model,omitempty"`
ResolvedModel string `json:"resolved_model,omitempty" bson:"resolved_model,omitempty"`
- Provider string `json:"provider" bson:"provider"`
+ Provider string `json:"provider" bson:"provider"` // canonical provider type used for routing and filters
+ ProviderName string `json:"provider_name,omitempty" bson:"provider_name,omitempty"`
AliasUsed bool `json:"alias_used,omitempty" bson:"alias_used,omitempty"`
ExecutionPlanVersionID string `json:"execution_plan_version_id,omitempty" bson:"execution_plan_version_id,omitempty"`
CacheType string `json:"cache_type,omitempty" bson:"cache_type,omitempty"`
@@ -143,6 +144,13 @@ func normalizeCacheType(value string) string {
}
}
+func displayAuditProviderName(providerName, provider string) string {
+ if trimmed := strings.TrimSpace(providerName); trimmed != "" {
+ return trimmed
+ }
+ return strings.TrimSpace(provider)
+}
+
// RedactedHeaders contains headers that should be automatically redacted.
// Values are replaced with "[REDACTED]" to prevent leaking secrets.
var RedactedHeaders = []string{
diff --git a/internal/auditlog/auditlog_test.go b/internal/auditlog/auditlog_test.go
index e2081c994..b9d083827 100644
--- a/internal/auditlog/auditlog_test.go
+++ b/internal/auditlog/auditlog_test.go
@@ -119,20 +119,20 @@ func TestRedactHeaders(t *testing.T) {
func TestLogEntryJSON(t *testing.T) {
entry := &LogEntry{
- ID: "test-id-123",
- Timestamp: time.Date(2024, 1, 15, 10, 30, 0, 0, time.UTC),
- DurationNs: 1500000,
- Model: "friendly-alias",
- ResolvedModel: "openai/gpt-4",
- Provider: "openai",
- AliasUsed: true,
- CacheType: CacheTypeExact,
- StatusCode: 200,
- RequestID: "req-123",
- ClientIP: "192.168.1.1",
- Method: "POST",
- Path: "/v1/chat/completions",
- Stream: false,
+ ID: "test-id-123",
+ Timestamp: time.Date(2024, 1, 15, 10, 30, 0, 0, time.UTC),
+ DurationNs: 1500000,
+ RequestedModel: "friendly-alias",
+ ResolvedModel: "openai/gpt-4",
+ Provider: "openai",
+ AliasUsed: true,
+ CacheType: CacheTypeExact,
+ StatusCode: 200,
+ RequestID: "req-123",
+ ClientIP: "192.168.1.1",
+ Method: "POST",
+ Path: "/v1/chat/completions",
+ Stream: false,
Data: &LogData{
UserAgent: "test-agent",
},
@@ -154,8 +154,8 @@ func TestLogEntryJSON(t *testing.T) {
if decoded.ID != entry.ID {
t.Errorf("ID mismatch: expected %q, got %q", entry.ID, decoded.ID)
}
- if decoded.Model != entry.Model {
- t.Errorf("Model mismatch: expected %q, got %q", entry.Model, decoded.Model)
+ if decoded.RequestedModel != entry.RequestedModel {
+ t.Errorf("RequestedModel mismatch: expected %q, got %q", entry.RequestedModel, decoded.RequestedModel)
}
if decoded.Provider != entry.Provider {
t.Errorf("Provider mismatch: expected %q, got %q", entry.Provider, decoded.Provider)
@@ -308,9 +308,9 @@ func TestLogger(t *testing.T) {
// Write some entries
for i := range 5 {
logger.Write(&LogEntry{
- ID: fmt.Sprintf("entry-%d", i),
- Timestamp: time.Now(),
- Model: "test-model",
+ ID: fmt.Sprintf("entry-%d", i),
+ Timestamp: time.Now(),
+ RequestedModel: "test-model",
})
}
@@ -498,8 +498,8 @@ func TestMiddleware_PrefersExecutionPlanOverLegacyResolution(t *testing.T) {
}
entry := logger.entries[0]
- if entry.Model != "anthropic/claude-opus-4-6" {
- t.Fatalf("Model = %q, want requested alias", entry.Model)
+ if entry.RequestedModel != "anthropic/claude-opus-4-6" {
+ t.Fatalf("RequestedModel = %q, want requested alias", entry.RequestedModel)
}
if entry.ResolvedModel != "openai/gpt-5-nano" {
t.Fatalf("ResolvedModel = %q, want openai/gpt-5-nano", entry.ResolvedModel)
@@ -573,8 +573,8 @@ func TestMiddleware_DoesNotApplyModelMetadataWithoutExecutionPlan(t *testing.T)
}
entry := logger.entries[0]
- if entry.Model != "" {
- t.Fatalf("Model = %q, want empty", entry.Model)
+ if entry.RequestedModel != "" {
+ t.Fatalf("RequestedModel = %q, want empty", entry.RequestedModel)
}
if entry.ResolvedModel != "" {
t.Fatalf("ResolvedModel = %q, want empty", entry.ResolvedModel)
@@ -619,8 +619,8 @@ func TestMiddleware_PassthroughExecutionPlanUsesPassthroughModel(t *testing.T) {
}
entry := logger.entries[0]
- if entry.Model != "gpt-4.1-nano" {
- t.Fatalf("Model = %q, want gpt-4.1-nano", entry.Model)
+ if entry.RequestedModel != "gpt-4.1-nano" {
+ t.Fatalf("RequestedModel = %q, want gpt-4.1-nano", entry.RequestedModel)
}
if entry.Provider != "openai" {
t.Fatalf("Provider = %q, want openai", entry.Provider)
@@ -890,10 +890,10 @@ data: [DONE]
logger := NewLogger(store, cfg)
entry := &LogEntry{
- ID: "test-entry",
- Timestamp: time.Now(),
- Model: "gpt-4",
- Data: &LogData{},
+ ID: "test-entry",
+ Timestamp: time.Now(),
+ RequestedModel: "gpt-4",
+ Data: &LogData{},
}
observedStream := streaming.NewObservedSSEStream(
@@ -943,7 +943,7 @@ func TestCreateStreamEntry(t *testing.T) {
ID: "test-id",
Timestamp: time.Now(),
DurationNs: 1000,
- Model: "claude-opus-4-6",
+ RequestedModel: "claude-opus-4-6",
ResolvedModel: "openai/gpt-5-nano",
Provider: "openai",
AliasUsed: true,
@@ -983,8 +983,8 @@ func TestCreateStreamEntry(t *testing.T) {
if streamEntry.ID != baseEntry.ID {
t.Errorf("ID mismatch")
}
- if streamEntry.Model != baseEntry.Model {
- t.Errorf("Model mismatch")
+ if streamEntry.RequestedModel != baseEntry.RequestedModel {
+ t.Errorf("RequestedModel mismatch")
}
if streamEntry.ResolvedModel != baseEntry.ResolvedModel {
t.Errorf("ResolvedModel mismatch")
diff --git a/internal/auditlog/logger.go b/internal/auditlog/logger.go
index de50a9f8e..8be33369b 100644
--- a/internal/auditlog/logger.go
+++ b/internal/auditlog/logger.go
@@ -79,7 +79,7 @@ func (l *Logger) Write(entry *LogEntry) {
}
slog.Warn("audit log buffer full, dropping entry",
"request_id", requestID,
- "model", entry.Model,
+ "requested_model", entry.RequestedModel,
)
}
}
diff --git a/internal/auditlog/middleware.go b/internal/auditlog/middleware.go
index 465eedc7b..3b06be0f6 100644
--- a/internal/auditlog/middleware.go
+++ b/internal/auditlog/middleware.go
@@ -190,14 +190,14 @@ func enrichEntryWithExecutionPlan(entry *LogEntry, plan *core.ExecutionPlan) {
entry.RequestID = requestID
}
if requestedModel := plan.RequestedQualifiedModel(); requestedModel != "" {
- entry.Model = requestedModel
+ entry.RequestedModel = requestedModel
}
- if resolvedModel := plan.ResolvedQualifiedModel(); resolvedModel != "" {
+ if resolvedModel := resolvedModelForAuditLog(plan); resolvedModel != "" {
entry.ResolvedModel = resolvedModel
}
if plan.Mode == core.ExecutionModePassthrough && plan.Passthrough != nil {
if model := strings.TrimSpace(plan.Passthrough.Model); model != "" {
- entry.Model = model
+ entry.RequestedModel = model
}
}
if providerType := strings.TrimSpace(plan.ProviderType); providerType != "" {
@@ -206,6 +206,9 @@ func enrichEntryWithExecutionPlan(entry *LogEntry, plan *core.ExecutionPlan) {
entry.Provider = strings.TrimSpace(plan.Resolution.ProviderType)
}
if plan.Resolution != nil {
+ if providerName := strings.TrimSpace(plan.Resolution.ProviderName); providerName != "" {
+ entry.ProviderName = providerName
+ }
entry.AliasUsed = plan.Resolution.AliasApplied
}
if versionID := strings.TrimSpace(plan.ExecutionPlanVersionID()); versionID != "" {
@@ -222,6 +225,23 @@ func enrichEntryWithExecutionPlan(entry *LogEntry, plan *core.ExecutionPlan) {
}
}
+func resolvedModelForAuditLog(plan *core.ExecutionPlan) string {
+ if plan == nil || plan.Resolution == nil {
+ return ""
+ }
+ model := strings.TrimSpace(plan.Resolution.ResolvedSelector.Model)
+ if model == "" {
+ return ""
+ }
+ if providerName := strings.TrimSpace(plan.Resolution.ProviderName); providerName != "" {
+ return providerName + "/" + model
+ }
+ if provider := strings.TrimSpace(plan.Resolution.ResolvedSelector.Provider); provider != "" {
+ return provider + "/" + model
+ }
+ return model
+}
+
func captureLoggedRequestBody(entry *LogEntry, bodyBytes []byte) {
entry.Data.RequestBody = captureLoggedBody(bodyBytes)
}
@@ -364,7 +384,7 @@ func EnrichEntry(c *echo.Context, model, provider string) {
return
}
- entry.Model = model
+ entry.RequestedModel = model
entry.Provider = provider
}
@@ -392,6 +412,43 @@ func EnrichLogEntryWithExecutionPlan(entry *LogEntry, plan *core.ExecutionPlan)
enrichEntryWithExecutionPlan(entry, plan)
}
+// EnrichEntryWithResolvedRoute attaches the final executed route to the live
+// audit entry after execution resolved to a concrete provider/model.
+func EnrichEntryWithResolvedRoute(c *echo.Context, resolvedModel, providerType, providerName string) {
+ entryVal := c.Get(string(LogEntryKey))
+ if entryVal == nil {
+ return
+ }
+
+ entry, ok := entryVal.(*LogEntry)
+ if !ok || entry == nil {
+ return
+ }
+
+ enrichEntryWithResolvedRoute(entry, resolvedModel, providerType, providerName)
+}
+
+// EnrichLogEntryWithResolvedRoute attaches the final executed route directly to
+// an existing audit log entry.
+func EnrichLogEntryWithResolvedRoute(entry *LogEntry, resolvedModel, providerType, providerName string) {
+ enrichEntryWithResolvedRoute(entry, resolvedModel, providerType, providerName)
+}
+
+func enrichEntryWithResolvedRoute(entry *LogEntry, resolvedModel, providerType, providerName string) {
+ if entry == nil {
+ return
+ }
+ if resolvedModel = strings.TrimSpace(resolvedModel); resolvedModel != "" {
+ entry.ResolvedModel = resolvedModel
+ }
+ if providerType = strings.TrimSpace(providerType); providerType != "" {
+ entry.Provider = providerType
+ }
+ if providerName = strings.TrimSpace(providerName); providerName != "" {
+ entry.ProviderName = providerName
+ }
+}
+
// EnrichEntryWithCacheType attaches cache-hit metadata to the live audit entry.
// The value is intentionally sourced directly from the cache middleware, not
// inferred from response headers after the fact.
diff --git a/internal/auditlog/middleware_test.go b/internal/auditlog/middleware_test.go
new file mode 100644
index 000000000..fc25a7436
--- /dev/null
+++ b/internal/auditlog/middleware_test.go
@@ -0,0 +1,42 @@
+package auditlog
+
+import (
+ "net/http"
+ "net/http/httptest"
+ "testing"
+
+ "github.com/labstack/echo/v5"
+
+ "gomodel/internal/core"
+)
+
+func TestEnrichEntryWithExecutionPlan_PrefersProviderNameForResolvedModel(t *testing.T) {
+ e := echo.New()
+ req := httptest.NewRequest(http.MethodGet, "/", nil)
+ rec := httptest.NewRecorder()
+ c := e.NewContext(req, rec)
+
+ entry := &LogEntry{ID: "provider-name-prefill"}
+ c.Set(string(LogEntryKey), entry)
+
+ EnrichEntryWithExecutionPlan(c, &core.ExecutionPlan{
+ ProviderType: "openai",
+ Resolution: &core.RequestModelResolution{
+ ResolvedSelector: core.ModelSelector{
+ Provider: "openai",
+ Model: "gpt-5-nano",
+ },
+ ProviderName: "openai_test",
+ },
+ })
+
+ if got := entry.Provider; got != "openai" {
+ t.Fatalf("Provider = %q, want %q", got, "openai")
+ }
+ if got := entry.ProviderName; got != "openai_test" {
+ t.Fatalf("ProviderName = %q, want %q", got, "openai_test")
+ }
+ if got := entry.ResolvedModel; got != "openai_test/gpt-5-nano" {
+ t.Fatalf("ResolvedModel = %q, want %q", got, "openai_test/gpt-5-nano")
+ }
+}
diff --git a/internal/auditlog/reader.go b/internal/auditlog/reader.go
index 31d1bab53..0d0bac72a 100644
--- a/internal/auditlog/reader.go
+++ b/internal/auditlog/reader.go
@@ -14,17 +14,17 @@ type QueryParams struct {
// LogQueryParams specifies query parameters for paginated audit log retrieval.
type LogQueryParams struct {
QueryParams
- Model string
- Provider string
- Method string
- Path string
- UserPath string
- ErrorType string
- Search string
- StatusCode *int
- Stream *bool
- Limit int
- Offset int
+ RequestedModel string
+ Provider string // filter by provider name or provider type
+ Method string
+ Path string
+ UserPath string
+ ErrorType string
+ Search string
+ StatusCode *int
+ Stream *bool
+ Limit int
+ Offset int
}
// LogListResult holds a paginated list of audit log entries.
diff --git a/internal/auditlog/reader_mongodb.go b/internal/auditlog/reader_mongodb.go
index edaec0ecc..0f3704533 100644
--- a/internal/auditlog/reader_mongodb.go
+++ b/internal/auditlog/reader_mongodb.go
@@ -22,9 +22,11 @@ type mongoLogRow struct {
ID string `bson:"_id"`
Timestamp time.Time `bson:"timestamp"`
DurationNs int64 `bson:"duration_ns"`
- Model string `bson:"model"`
+ RequestedModel string `bson:"requested_model"`
+ LegacyModel string `bson:"model"`
ResolvedModel string `bson:"resolved_model"`
Provider string `bson:"provider"`
+ ProviderName string `bson:"provider_name"`
AliasUsed bool `bson:"alias_used"`
ExecutionPlanVersionID string `bson:"execution_plan_version_id"`
CacheType string `bson:"cache_type"`
@@ -46,9 +48,10 @@ func (r mongoLogRow) toLogEntry() *LogEntry {
ID: r.ID,
Timestamp: r.Timestamp,
DurationNs: r.DurationNs,
- Model: r.Model,
+ RequestedModel: firstNonEmpty(r.RequestedModel, r.LegacyModel),
ResolvedModel: r.ResolvedModel,
Provider: r.Provider,
+ ProviderName: displayAuditProviderName(r.ProviderName, r.Provider),
AliasUsed: r.AliasUsed,
ExecutionPlanVersionID: r.ExecutionPlanVersionID,
CacheType: normalizeCacheType(r.CacheType),
@@ -113,23 +116,30 @@ func (r *MongoDBReader) GetLogs(ctx context.Context, params LogQueryParams) (*Lo
if tsFilter := mongoDateRangeFilter(params.QueryParams); tsFilter != nil {
matchFilters = append(matchFilters, bson.E{Key: "timestamp", Value: tsFilter})
}
- if params.Model != "" {
+ if params.RequestedModel != "" {
matchFilters = append(matchFilters, bson.E{
- Key: "model",
- Value: bson.D{
- {Key: "$regex", Value: regexp.QuoteMeta(params.Model)},
- {Key: "$options", Value: "i"},
+ Key: "$or",
+ Value: bson.A{
+ bson.D{{Key: "requested_model", Value: bson.D{
+ {Key: "$regex", Value: regexp.QuoteMeta(params.RequestedModel)},
+ {Key: "$options", Value: "i"},
+ }}},
+ bson.D{{Key: "model", Value: bson.D{
+ {Key: "$regex", Value: regexp.QuoteMeta(params.RequestedModel)},
+ {Key: "$options", Value: "i"},
+ }}},
},
})
}
if params.Provider != "" {
- matchFilters = append(matchFilters, bson.E{
- Key: "provider",
- Value: bson.D{
- {Key: "$regex", Value: regexp.QuoteMeta(params.Provider)},
- {Key: "$options", Value: "i"},
- },
- })
+ regex := bson.D{
+ {Key: "$regex", Value: regexp.QuoteMeta(params.Provider)},
+ {Key: "$options", Value: "i"},
+ }
+ matchFilters = append(matchFilters, bson.E{Key: "$or", Value: bson.A{
+ bson.D{{Key: "provider", Value: regex}},
+ bson.D{{Key: "provider_name", Value: regex}},
+ }})
}
if params.Method != "" {
matchFilters = append(matchFilters, bson.E{Key: "method", Value: params.Method})
@@ -169,10 +179,13 @@ func (r *MongoDBReader) GetLogs(ctx context.Context, params LogQueryParams) (*Lo
matchFilters = append(matchFilters, bson.E{Key: "$or", Value: bson.A{
bson.D{{Key: "request_id", Value: regex}},
bson.D{{Key: "auth_key_id", Value: regex}},
+ bson.D{{Key: "requested_model", Value: regex}},
bson.D{{Key: "model", Value: regex}},
bson.D{{Key: "provider", Value: regex}},
+ bson.D{{Key: "provider_name", Value: regex}},
bson.D{{Key: "method", Value: regex}},
bson.D{{Key: "path", Value: regex}},
+ bson.D{{Key: "user_path", Value: regex}},
bson.D{{Key: "error_type", Value: regex}},
bson.D{{Key: "data.error_message", Value: regex}},
}})
@@ -238,6 +251,15 @@ func (r *MongoDBReader) GetLogs(ctx context.Context, params LogQueryParams) (*Lo
}, nil
}
+func firstNonEmpty(values ...string) string {
+ for _, value := range values {
+ if value != "" {
+ return value
+ }
+ }
+ return ""
+}
+
// GetLogByID returns a single audit log entry by ID.
func (r *MongoDBReader) GetLogByID(ctx context.Context, id string) (*LogEntry, error) {
var row mongoLogRow
diff --git a/internal/auditlog/reader_mongodb_test.go b/internal/auditlog/reader_mongodb_test.go
index 96dcc50af..bbd94fb0a 100644
--- a/internal/auditlog/reader_mongodb_test.go
+++ b/internal/auditlog/reader_mongodb_test.go
@@ -59,10 +59,10 @@ func TestSanitizeLogDataNilSafe(t *testing.T) {
func TestMongoLogRowToLogEntryPreservesCacheType(t *testing.T) {
row := mongoLogRow{
- ID: "log-1",
- Model: "gpt-4",
- Provider: "openai",
- CacheType: CacheTypeSemantic,
+ ID: "log-1",
+ RequestedModel: "gpt-4",
+ Provider: "openai",
+ CacheType: CacheTypeSemantic,
}
entry := row.toLogEntry()
diff --git a/internal/auditlog/reader_postgresql.go b/internal/auditlog/reader_postgresql.go
index 938c22af2..cb21d0b25 100644
--- a/internal/auditlog/reader_postgresql.go
+++ b/internal/auditlog/reader_postgresql.go
@@ -32,15 +32,15 @@ func (r *PostgreSQLReader) GetLogs(ctx context.Context, params LogQueryParams) (
return nil, err
}
- if params.Model != "" {
- conditions = append(conditions, fmt.Sprintf("model ILIKE $%d ESCAPE '\\'", argIdx))
- args = append(args, "%"+escapeLikeWildcards(params.Model)+"%")
+ if params.RequestedModel != "" {
+ conditions = append(conditions, fmt.Sprintf("requested_model ILIKE $%d ESCAPE '\\'", argIdx))
+ args = append(args, "%"+escapeLikeWildcards(params.RequestedModel)+"%")
argIdx++
}
if params.Provider != "" {
- conditions = append(conditions, fmt.Sprintf("provider ILIKE $%d ESCAPE '\\'", argIdx))
- args = append(args, "%"+escapeLikeWildcards(params.Provider)+"%")
- argIdx++
+ conditions = append(conditions, fmt.Sprintf("(provider ILIKE $%d ESCAPE '\\' OR provider_name ILIKE $%d ESCAPE '\\')", argIdx, argIdx+1))
+ args = append(args, "%"+escapeLikeWildcards(params.Provider)+"%", "%"+escapeLikeWildcards(params.Provider)+"%")
+ argIdx += 2
}
if params.Method != "" {
conditions = append(conditions, fmt.Sprintf("method = $%d", argIdx))
@@ -78,7 +78,7 @@ func (r *PostgreSQLReader) GetLogs(ctx context.Context, params LogQueryParams) (
}
if params.Search != "" {
s := "%" + escapeLikeWildcards(params.Search) + "%"
- conditions = append(conditions, fmt.Sprintf("(request_id ILIKE $%d ESCAPE '\\' OR auth_key_id ILIKE $%d ESCAPE '\\' OR model ILIKE $%d ESCAPE '\\' OR provider ILIKE $%d ESCAPE '\\' OR method ILIKE $%d ESCAPE '\\' OR path ILIKE $%d ESCAPE '\\' OR error_type ILIKE $%d ESCAPE '\\')", argIdx, argIdx, argIdx, argIdx, argIdx, argIdx, argIdx))
+ conditions = append(conditions, fmt.Sprintf("(request_id ILIKE $%d ESCAPE '\\' OR auth_key_id ILIKE $%d ESCAPE '\\' OR requested_model ILIKE $%d ESCAPE '\\' OR provider ILIKE $%d ESCAPE '\\' OR provider_name ILIKE $%d ESCAPE '\\' OR method ILIKE $%d ESCAPE '\\' OR path ILIKE $%d ESCAPE '\\' OR user_path ILIKE $%d ESCAPE '\\' OR error_type ILIKE $%d ESCAPE '\\')", argIdx, argIdx, argIdx, argIdx, argIdx, argIdx, argIdx, argIdx, argIdx))
args = append(args, s)
argIdx++
}
@@ -91,7 +91,7 @@ func (r *PostgreSQLReader) GetLogs(ctx context.Context, params LogQueryParams) (
return nil, fmt.Errorf("failed to count audit log entries: %w", err)
}
- dataQuery := fmt.Sprintf(`SELECT id, timestamp, duration_ns, model, resolved_model, provider, alias_used, execution_plan_version_id, cache_type, status_code, request_id, auth_key_id, auth_method,
+ dataQuery := fmt.Sprintf(`SELECT id, timestamp, duration_ns, requested_model, resolved_model, provider, provider_name, alias_used, execution_plan_version_id, cache_type, status_code, request_id, auth_key_id, auth_method,
client_ip, method, path, user_path, stream, error_type, data
FROM audit_logs%s ORDER BY timestamp DESC LIMIT $%d OFFSET $%d`, where, argIdx, argIdx+1)
dataArgs := append(append([]any(nil), args...), limit, offset)
@@ -106,13 +106,14 @@ func (r *PostgreSQLReader) GetLogs(ctx context.Context, params LogQueryParams) (
for rows.Next() {
var e LogEntry
var dataJSON *string
+ var providerName *string
var executionPlanVersionID *string
var cacheType *string
var authKeyID *string
var authMethod *string
var userPath *string
- if err := rows.Scan(&e.ID, &e.Timestamp, &e.DurationNs, &e.Model, &e.ResolvedModel, &e.Provider, &e.AliasUsed, &executionPlanVersionID, &cacheType, &e.StatusCode,
+ if err := rows.Scan(&e.ID, &e.Timestamp, &e.DurationNs, &e.RequestedModel, &e.ResolvedModel, &e.Provider, &providerName, &e.AliasUsed, &executionPlanVersionID, &cacheType, &e.StatusCode,
&e.RequestID, &authKeyID, &authMethod, &e.ClientIP, &e.Method, &e.Path, &userPath, &e.Stream, &e.ErrorType, &dataJSON); err != nil {
return nil, fmt.Errorf("failed to scan audit log row: %w", err)
}
@@ -128,6 +129,11 @@ func (r *PostgreSQLReader) GetLogs(ctx context.Context, params LogQueryParams) (
if cacheType != nil {
e.CacheType = normalizeCacheType(*cacheType)
}
+ if providerName != nil {
+ e.ProviderName = displayAuditProviderName(*providerName, e.Provider)
+ } else {
+ e.ProviderName = displayAuditProviderName("", e.Provider)
+ }
if userPath != nil {
e.UserPath = *userPath
}
@@ -158,7 +164,7 @@ func (r *PostgreSQLReader) GetLogs(ctx context.Context, params LogQueryParams) (
// GetLogByID returns a single audit log entry by ID.
func (r *PostgreSQLReader) GetLogByID(ctx context.Context, id string) (*LogEntry, error) {
- query := `SELECT id, timestamp, duration_ns, model, resolved_model, provider, alias_used, execution_plan_version_id, cache_type, status_code, request_id, auth_key_id, auth_method,
+ query := `SELECT id, timestamp, duration_ns, requested_model, resolved_model, provider, provider_name, alias_used, execution_plan_version_id, cache_type, status_code, request_id, auth_key_id, auth_method,
client_ip, method, path, user_path, stream, error_type, data
FROM audit_logs WHERE id::text = $1 LIMIT 1`
@@ -200,7 +206,7 @@ func pgDateRangeConditions(params QueryParams, argIdx int) (conditions []string,
}
func (r *PostgreSQLReader) findByResponseID(ctx context.Context, responseID string) (*LogEntry, error) {
- query := `SELECT id, timestamp, duration_ns, model, resolved_model, provider, alias_used, execution_plan_version_id, cache_type, status_code, request_id, auth_key_id, auth_method,
+ query := `SELECT id, timestamp, duration_ns, requested_model, resolved_model, provider, provider_name, alias_used, execution_plan_version_id, cache_type, status_code, request_id, auth_key_id, auth_method,
client_ip, method, path, user_path, stream, error_type, data
FROM audit_logs
WHERE data->'response_body'->>'id' = $1
@@ -219,7 +225,7 @@ func (r *PostgreSQLReader) findByResponseID(ctx context.Context, responseID stri
}
func (r *PostgreSQLReader) findByPreviousResponseID(ctx context.Context, previousResponseID string) (*LogEntry, error) {
- query := `SELECT id, timestamp, duration_ns, model, resolved_model, provider, alias_used, execution_plan_version_id, cache_type, status_code, request_id, auth_key_id, auth_method,
+ query := `SELECT id, timestamp, duration_ns, requested_model, resolved_model, provider, provider_name, alias_used, execution_plan_version_id, cache_type, status_code, request_id, auth_key_id, auth_method,
client_ip, method, path, user_path, stream, error_type, data
FROM audit_logs
WHERE data->'request_body'->>'previous_response_id' = $1
@@ -242,13 +248,14 @@ func scanPostgreSQLLogEntry(rows interface {
}) (*LogEntry, error) {
var e LogEntry
var dataJSON *string
+ var providerName *string
var executionPlanVersionID *string
var cacheType *string
var authKeyID *string
var authMethod *string
var userPath *string
- if err := rows.Scan(&e.ID, &e.Timestamp, &e.DurationNs, &e.Model, &e.ResolvedModel, &e.Provider, &e.AliasUsed, &executionPlanVersionID, &cacheType, &e.StatusCode,
+ if err := rows.Scan(&e.ID, &e.Timestamp, &e.DurationNs, &e.RequestedModel, &e.ResolvedModel, &e.Provider, &providerName, &e.AliasUsed, &executionPlanVersionID, &cacheType, &e.StatusCode,
&e.RequestID, &authKeyID, &authMethod, &e.ClientIP, &e.Method, &e.Path, &userPath, &e.Stream, &e.ErrorType, &dataJSON); err != nil {
return nil, fmt.Errorf("failed to scan audit log row: %w", err)
}
@@ -264,6 +271,11 @@ func scanPostgreSQLLogEntry(rows interface {
if cacheType != nil {
e.CacheType = normalizeCacheType(*cacheType)
}
+ if providerName != nil {
+ e.ProviderName = displayAuditProviderName(*providerName, e.Provider)
+ } else {
+ e.ProviderName = displayAuditProviderName("", e.Provider)
+ }
if userPath != nil {
e.UserPath = *userPath
}
diff --git a/internal/auditlog/reader_sqlite.go b/internal/auditlog/reader_sqlite.go
index 4f7162fa5..a18fec257 100644
--- a/internal/auditlog/reader_sqlite.go
+++ b/internal/auditlog/reader_sqlite.go
@@ -35,13 +35,13 @@ func (r *SQLiteReader) GetLogs(ctx context.Context, params LogQueryParams) (*Log
return nil, err
}
- if params.Model != "" {
- conditions = append(conditions, "model LIKE ? ESCAPE '\\'")
- args = append(args, "%"+escapeLikeWildcards(params.Model)+"%")
+ if params.RequestedModel != "" {
+ conditions = append(conditions, "requested_model LIKE ? ESCAPE '\\'")
+ args = append(args, "%"+escapeLikeWildcards(params.RequestedModel)+"%")
}
if params.Provider != "" {
- conditions = append(conditions, "provider LIKE ? ESCAPE '\\'")
- args = append(args, "%"+escapeLikeWildcards(params.Provider)+"%")
+ conditions = append(conditions, "(provider LIKE ? ESCAPE '\\' OR provider_name LIKE ? ESCAPE '\\')")
+ args = append(args, "%"+escapeLikeWildcards(params.Provider)+"%", "%"+escapeLikeWildcards(params.Provider)+"%")
}
if params.Method != "" {
conditions = append(conditions, "method = ?")
@@ -73,8 +73,8 @@ func (r *SQLiteReader) GetLogs(ctx context.Context, params LogQueryParams) (*Log
}
if params.Search != "" {
s := "%" + escapeLikeWildcards(params.Search) + "%"
- conditions = append(conditions, `(request_id LIKE ? ESCAPE '\' OR auth_key_id LIKE ? ESCAPE '\' OR model LIKE ? ESCAPE '\' OR provider LIKE ? ESCAPE '\' OR method LIKE ? ESCAPE '\' OR path LIKE ? ESCAPE '\' OR error_type LIKE ? ESCAPE '\')`)
- args = append(args, s, s, s, s, s, s, s)
+ conditions = append(conditions, `(request_id LIKE ? ESCAPE '\' OR auth_key_id LIKE ? ESCAPE '\' OR requested_model LIKE ? ESCAPE '\' OR provider LIKE ? ESCAPE '\' OR provider_name LIKE ? ESCAPE '\' OR method LIKE ? ESCAPE '\' OR path LIKE ? ESCAPE '\' OR user_path LIKE ? ESCAPE '\' OR error_type LIKE ? ESCAPE '\')`)
+ args = append(args, s, s, s, s, s, s, s, s, s)
}
where := buildWhereClause(conditions)
@@ -86,7 +86,7 @@ func (r *SQLiteReader) GetLogs(ctx context.Context, params LogQueryParams) (*Log
return nil, fmt.Errorf("failed to count audit log entries: %w", err)
}
- dataQuery := `SELECT id, timestamp, duration_ns, model, resolved_model, provider, alias_used, execution_plan_version_id, cache_type, status_code, request_id, auth_key_id, auth_method,
+ dataQuery := `SELECT id, timestamp, duration_ns, requested_model, resolved_model, provider, provider_name, alias_used, execution_plan_version_id, cache_type, status_code, request_id, auth_key_id, auth_method,
client_ip, method, path, user_path, stream, error_type, data
FROM audit_logs` + where + ` ORDER BY timestamp DESC LIMIT ? OFFSET ?`
dataArgs := append(append([]any(nil), args...), limit, offset)
@@ -101,6 +101,7 @@ func (r *SQLiteReader) GetLogs(ctx context.Context, params LogQueryParams) (*Log
for rows.Next() {
var e LogEntry
var ts string
+ var providerName sql.NullString
var aliasUsedInt int
var streamInt int
var dataJSON *string
@@ -110,7 +111,7 @@ func (r *SQLiteReader) GetLogs(ctx context.Context, params LogQueryParams) (*Log
var authMethod sql.NullString
var userPath sql.NullString
- if err := rows.Scan(&e.ID, &ts, &e.DurationNs, &e.Model, &e.ResolvedModel, &e.Provider, &aliasUsedInt, &executionPlanVersionID, &cacheType, &e.StatusCode,
+ if err := rows.Scan(&e.ID, &ts, &e.DurationNs, &e.RequestedModel, &e.ResolvedModel, &e.Provider, &providerName, &aliasUsedInt, &executionPlanVersionID, &cacheType, &e.StatusCode,
&e.RequestID, &authKeyID, &authMethod, &e.ClientIP, &e.Method, &e.Path, &userPath, &streamInt, &e.ErrorType, &dataJSON); err != nil {
return nil, fmt.Errorf("failed to scan audit log row: %w", err)
}
@@ -130,6 +131,11 @@ func (r *SQLiteReader) GetLogs(ctx context.Context, params LogQueryParams) (*Log
if cacheType.Valid {
e.CacheType = normalizeCacheType(cacheType.String)
}
+ if providerName.Valid {
+ e.ProviderName = displayAuditProviderName(providerName.String, e.Provider)
+ } else {
+ e.ProviderName = displayAuditProviderName("", e.Provider)
+ }
if userPath.Valid {
e.UserPath = userPath.String
}
@@ -160,7 +166,7 @@ func (r *SQLiteReader) GetLogs(ctx context.Context, params LogQueryParams) (*Log
// GetLogByID returns a single audit log entry by ID.
func (r *SQLiteReader) GetLogByID(ctx context.Context, id string) (*LogEntry, error) {
- query := `SELECT id, timestamp, duration_ns, model, resolved_model, provider, alias_used, execution_plan_version_id, cache_type, status_code, request_id, auth_key_id, auth_method,
+ query := `SELECT id, timestamp, duration_ns, requested_model, resolved_model, provider, provider_name, alias_used, execution_plan_version_id, cache_type, status_code, request_id, auth_key_id, auth_method,
client_ip, method, path, user_path, stream, error_type, data
FROM audit_logs WHERE id = ? LIMIT 1`
@@ -292,7 +298,7 @@ func parseSQLTimestamp(ts string, entryID string) time.Time {
}
func (r *SQLiteReader) findByResponseID(ctx context.Context, responseID string) (*LogEntry, error) {
- query := `SELECT id, timestamp, duration_ns, model, resolved_model, provider, alias_used, execution_plan_version_id, cache_type, status_code, request_id, auth_key_id, auth_method,
+ query := `SELECT id, timestamp, duration_ns, requested_model, resolved_model, provider, provider_name, alias_used, execution_plan_version_id, cache_type, status_code, request_id, auth_key_id, auth_method,
client_ip, method, path, user_path, stream, error_type, data
FROM audit_logs
WHERE json_extract(data, '$.response_body.id') = ?
@@ -311,7 +317,7 @@ func (r *SQLiteReader) findByResponseID(ctx context.Context, responseID string)
}
func (r *SQLiteReader) findByPreviousResponseID(ctx context.Context, previousResponseID string) (*LogEntry, error) {
- query := `SELECT id, timestamp, duration_ns, model, resolved_model, provider, alias_used, execution_plan_version_id, cache_type, status_code, request_id, auth_key_id, auth_method,
+ query := `SELECT id, timestamp, duration_ns, requested_model, resolved_model, provider, provider_name, alias_used, execution_plan_version_id, cache_type, status_code, request_id, auth_key_id, auth_method,
client_ip, method, path, user_path, stream, error_type, data
FROM audit_logs
WHERE json_extract(data, '$.request_body.previous_response_id') = ?
@@ -332,6 +338,7 @@ func (r *SQLiteReader) findByPreviousResponseID(ctx context.Context, previousRes
func scanSQLiteLogEntry(rows *sql.Rows) (*LogEntry, error) {
var e LogEntry
var ts string
+ var providerName sql.NullString
var aliasUsedInt int
var streamInt int
var dataJSON *string
@@ -341,7 +348,7 @@ func scanSQLiteLogEntry(rows *sql.Rows) (*LogEntry, error) {
var authMethod sql.NullString
var userPath sql.NullString
- if err := rows.Scan(&e.ID, &ts, &e.DurationNs, &e.Model, &e.ResolvedModel, &e.Provider, &aliasUsedInt, &executionPlanVersionID, &cacheType, &e.StatusCode,
+ if err := rows.Scan(&e.ID, &ts, &e.DurationNs, &e.RequestedModel, &e.ResolvedModel, &e.Provider, &providerName, &aliasUsedInt, &executionPlanVersionID, &cacheType, &e.StatusCode,
&e.RequestID, &authKeyID, &authMethod, &e.ClientIP, &e.Method, &e.Path, &userPath, &streamInt, &e.ErrorType, &dataJSON); err != nil {
return nil, fmt.Errorf("failed to scan audit log row: %w", err)
}
@@ -361,6 +368,11 @@ func scanSQLiteLogEntry(rows *sql.Rows) (*LogEntry, error) {
if cacheType.Valid {
e.CacheType = normalizeCacheType(cacheType.String)
}
+ if providerName.Valid {
+ e.ProviderName = displayAuditProviderName(providerName.String, e.Provider)
+ } else {
+ e.ProviderName = displayAuditProviderName("", e.Provider)
+ }
if userPath.Valid {
e.UserPath = userPath.String
}
diff --git a/internal/auditlog/reader_sqlite_boundary_test.go b/internal/auditlog/reader_sqlite_boundary_test.go
index 8ae536093..9686a9495 100644
--- a/internal/auditlog/reader_sqlite_boundary_test.go
+++ b/internal/auditlog/reader_sqlite_boundary_test.go
@@ -18,22 +18,22 @@ func TestSQLiteReaderGetLogs_IncludesFractionalStartBoundaryAndExcludesFractiona
ctx := context.Background()
err = store.WriteBatch(ctx, []*LogEntry{
{
- ID: "start-boundary",
- Timestamp: time.Date(2026, 1, 15, 23, 0, 0, 123_000_000, time.UTC),
- Model: "gpt-5",
- Provider: "openai",
+ ID: "start-boundary",
+ Timestamp: time.Date(2026, 1, 15, 23, 0, 0, 123_000_000, time.UTC),
+ RequestedModel: "gpt-5",
+ Provider: "openai",
},
{
- ID: "inside-range",
- Timestamp: time.Date(2026, 1, 16, 12, 0, 0, 0, time.UTC),
- Model: "gpt-5",
- Provider: "openai",
+ ID: "inside-range",
+ Timestamp: time.Date(2026, 1, 16, 12, 0, 0, 0, time.UTC),
+ RequestedModel: "gpt-5",
+ Provider: "openai",
},
{
- ID: "after-end-boundary",
- Timestamp: time.Date(2026, 1, 16, 23, 0, 0, 123_000_000, time.UTC),
- Model: "gpt-5",
- Provider: "openai",
+ ID: "after-end-boundary",
+ Timestamp: time.Date(2026, 1, 16, 23, 0, 0, 123_000_000, time.UTC),
+ RequestedModel: "gpt-5",
+ Provider: "openai",
},
})
if err != nil {
@@ -75,3 +75,56 @@ func TestSQLiteReaderGetLogs_IncludesFractionalStartBoundaryAndExcludesFractiona
t.Fatalf("expected boundary entry %q, got %q", "start-boundary", result.Entries[1].ID)
}
}
+
+func TestSQLiteReaderGetLogs_SearchMatchesUserPath(t *testing.T) {
+ db := createTestDB(t)
+ defer db.Close()
+
+ store, err := NewSQLiteStore(db, 0)
+ if err != nil {
+ t.Fatalf("failed to create store: %v", err)
+ }
+
+ ctx := context.Background()
+ if err := store.WriteBatch(ctx, []*LogEntry{
+ {
+ ID: "team-match",
+ Timestamp: time.Date(2026, 1, 16, 12, 0, 0, 0, time.UTC),
+ RequestedModel: "gpt-5",
+ Provider: "openai",
+ UserPath: "/team/alpha",
+ },
+ {
+ ID: "other-team",
+ Timestamp: time.Date(2026, 1, 16, 11, 0, 0, 0, time.UTC),
+ RequestedModel: "gpt-5",
+ Provider: "openai",
+ UserPath: "/org/beta",
+ },
+ }); err != nil {
+ t.Fatalf("failed to seed audit logs: %v", err)
+ }
+
+ reader, err := NewSQLiteReader(db)
+ if err != nil {
+ t.Fatalf("failed to create reader: %v", err)
+ }
+
+ result, err := reader.GetLogs(ctx, LogQueryParams{
+ Search: "team/alpha",
+ Limit: 10,
+ })
+ if err != nil {
+ t.Fatalf("GetLogs returned error: %v", err)
+ }
+
+ if result.Total != 1 {
+ t.Fatalf("expected 1 log in search result, got %d", result.Total)
+ }
+ if len(result.Entries) != 1 {
+ t.Fatalf("expected 1 returned entry, got %d", len(result.Entries))
+ }
+ if result.Entries[0].ID != "team-match" {
+ t.Fatalf("expected matching entry %q, got %q", "team-match", result.Entries[0].ID)
+ }
+}
diff --git a/internal/auditlog/store_mongodb.go b/internal/auditlog/store_mongodb.go
index fbe812e12..ad1d02cb6 100644
--- a/internal/auditlog/store_mongodb.go
+++ b/internal/auditlog/store_mongodb.go
@@ -65,7 +65,7 @@ func NewMongoDBStore(database *mongo.Database, retentionDays int) (*MongoDBStore
// Create indexes for common queries
indexes := []mongo.IndexModel{
{
- Keys: bson.D{{Key: "model", Value: 1}},
+ Keys: bson.D{{Key: "requested_model", Value: 1}},
},
{
Keys: bson.D{{Key: "status_code", Value: 1}},
@@ -73,6 +73,9 @@ func NewMongoDBStore(database *mongo.Database, retentionDays int) (*MongoDBStore
{
Keys: bson.D{{Key: "provider", Value: 1}},
},
+ {
+ Keys: bson.D{{Key: "provider_name", Value: 1}},
+ },
{
Keys: bson.D{{Key: "execution_plan_version_id", Value: 1}},
},
diff --git a/internal/auditlog/store_postgresql.go b/internal/auditlog/store_postgresql.go
index 0fa0f1a7a..0b3e18621 100644
--- a/internal/auditlog/store_postgresql.go
+++ b/internal/auditlog/store_postgresql.go
@@ -14,13 +14,13 @@ import (
)
const (
- auditLogInsertColumnCount = 20
+ auditLogInsertColumnCount = 21
postgresMaxBindParameters = 65535
auditLogInsertMaxRowsPerQuery = postgresMaxBindParameters / auditLogInsertColumnCount
)
const auditLogInsertPrefix = `
- INSERT INTO audit_logs (id, timestamp, duration_ns, model, resolved_model, provider, alias_used, execution_plan_version_id, cache_type, status_code,
+ INSERT INTO audit_logs (id, timestamp, duration_ns, requested_model, resolved_model, provider, provider_name, alias_used, execution_plan_version_id, cache_type, status_code,
request_id, auth_key_id, auth_method, client_ip, method, path, user_path, stream, error_type, data)
VALUES `
@@ -56,9 +56,10 @@ func NewPostgreSQLStore(pool *pgxpool.Pool, retentionDays int) (*PostgreSQLStore
id UUID PRIMARY KEY,
timestamp TIMESTAMPTZ NOT NULL,
duration_ns BIGINT DEFAULT 0,
- model TEXT,
+ requested_model TEXT,
resolved_model TEXT,
provider TEXT,
+ provider_name TEXT,
alias_used BOOLEAN DEFAULT FALSE,
execution_plan_version_id TEXT,
cache_type TEXT,
@@ -79,8 +80,14 @@ func NewPostgreSQLStore(pool *pgxpool.Pool, retentionDays int) (*PostgreSQLStore
return nil, fmt.Errorf("failed to create audit_logs table: %w", err)
}
+ if err := renamePostgreSQLAuditColumn(ctx, pool, "audit_logs", "model", "requested_model"); err != nil {
+ return nil, fmt.Errorf("failed to rename audit_logs.model to requested_model: %w", err)
+ }
+
migrations := []string{
+ "ALTER TABLE audit_logs ADD COLUMN IF NOT EXISTS requested_model TEXT",
"ALTER TABLE audit_logs ADD COLUMN IF NOT EXISTS resolved_model TEXT",
+ "ALTER TABLE audit_logs ADD COLUMN IF NOT EXISTS provider_name TEXT",
"ALTER TABLE audit_logs ADD COLUMN IF NOT EXISTS alias_used BOOLEAN DEFAULT FALSE",
"ALTER TABLE audit_logs ADD COLUMN IF NOT EXISTS execution_plan_version_id TEXT",
"ALTER TABLE audit_logs ADD COLUMN IF NOT EXISTS cache_type TEXT",
@@ -97,9 +104,11 @@ func NewPostgreSQLStore(pool *pgxpool.Pool, retentionDays int) (*PostgreSQLStore
// Create indexes for common queries
indexes := []string{
"CREATE INDEX IF NOT EXISTS idx_audit_timestamp ON audit_logs(timestamp)",
- "CREATE INDEX IF NOT EXISTS idx_audit_model ON audit_logs(model)",
+ "DROP INDEX IF EXISTS idx_audit_model",
+ "CREATE INDEX IF NOT EXISTS idx_audit_requested_model ON audit_logs(requested_model)",
"CREATE INDEX IF NOT EXISTS idx_audit_status ON audit_logs(status_code)",
"CREATE INDEX IF NOT EXISTS idx_audit_provider ON audit_logs(provider)",
+ "CREATE INDEX IF NOT EXISTS idx_audit_provider_name ON audit_logs(provider_name)",
"CREATE INDEX IF NOT EXISTS idx_audit_execution_plan_version_id ON audit_logs(execution_plan_version_id)",
"CREATE INDEX IF NOT EXISTS idx_audit_request_id ON audit_logs(request_id)",
"CREATE INDEX IF NOT EXISTS idx_audit_auth_key_id ON audit_logs(auth_key_id)",
@@ -222,9 +231,10 @@ func buildAuditLogInsert(entries []*LogEntry) (string, []any) {
entry.ID,
entry.Timestamp,
entry.DurationNs,
- entry.Model,
+ entry.RequestedModel,
entry.ResolvedModel,
entry.Provider,
+ entry.ProviderName,
entry.AliasUsed,
entry.ExecutionPlanVersionID,
cacheTypeValue,
@@ -246,6 +256,33 @@ func buildAuditLogInsert(entries []*LogEntry) (string, []any) {
return builder.String(), args
}
+func renamePostgreSQLAuditColumn(ctx context.Context, pool *pgxpool.Pool, tableName, from, to string) error {
+ fromExists, err := postgresqlColumnExists(ctx, pool, tableName, from)
+ if err != nil || !fromExists {
+ return err
+ }
+ toExists, err := postgresqlColumnExists(ctx, pool, tableName, to)
+ if err != nil || toExists {
+ return err
+ }
+ _, err = pool.Exec(ctx, fmt.Sprintf("ALTER TABLE %s RENAME COLUMN %s TO %s", tableName, from, to))
+ return err
+}
+
+func postgresqlColumnExists(ctx context.Context, pool *pgxpool.Pool, tableName, columnName string) (bool, error) {
+ var exists bool
+ err := pool.QueryRow(ctx, `
+ SELECT EXISTS (
+ SELECT 1
+ FROM information_schema.columns
+ WHERE table_schema = current_schema()
+ AND table_name = $1
+ AND column_name = $2
+ )
+ `, tableName, columnName).Scan(&exists)
+ return exists, err
+}
+
// Flush is a no-op for PostgreSQL as writes are synchronous.
func (s *PostgreSQLStore) Flush(_ context.Context) error {
return nil
diff --git a/internal/auditlog/store_postgresql_test.go b/internal/auditlog/store_postgresql_test.go
index 16d9f3161..35eb9f57f 100644
--- a/internal/auditlog/store_postgresql_test.go
+++ b/internal/auditlog/store_postgresql_test.go
@@ -11,97 +11,101 @@ func TestBuildAuditLogInsert(t *testing.T) {
query, args := buildAuditLogInsert([]*LogEntry{
{
- ID: "log-1",
- Timestamp: now,
- DurationNs: 1234,
- Model: "gpt-4o-mini",
- ResolvedModel: "gpt-4o-mini",
- Provider: "openai",
- AliasUsed: true,
- CacheType: CacheTypeExact,
- StatusCode: 200,
- RequestID: "req-1",
- AuthKeyID: "auth-key-1",
- ClientIP: "127.0.0.1",
- Method: "POST",
- Path: "/v1/chat/completions",
- UserPath: "/team/alpha",
- Stream: true,
- ErrorType: "",
+ ID: "log-1",
+ Timestamp: now,
+ DurationNs: 1234,
+ RequestedModel: "gpt-4o-mini",
+ ResolvedModel: "gpt-4o-mini",
+ Provider: "openai",
+ ProviderName: "primary-openai",
+ AliasUsed: true,
+ CacheType: CacheTypeExact,
+ StatusCode: 200,
+ RequestID: "req-1",
+ AuthKeyID: "auth-key-1",
+ ClientIP: "127.0.0.1",
+ Method: "POST",
+ Path: "/v1/chat/completions",
+ UserPath: "/team/alpha",
+ Stream: true,
+ ErrorType: "",
Data: &LogData{
UserAgent: "test-agent",
},
},
{
- ID: "log-2",
- Timestamp: now.Add(time.Second),
- DurationNs: 5678,
- Model: "gpt-4.1",
- ResolvedModel: "gpt-4.1",
- Provider: "openai",
- AliasUsed: false,
- StatusCode: 500,
- RequestID: "req-2",
- ClientIP: "10.0.0.1",
- Method: "POST",
- Path: "/v1/responses",
- Stream: false,
- ErrorType: "server_error",
- Data: nil,
+ ID: "log-2",
+ Timestamp: now.Add(time.Second),
+ DurationNs: 5678,
+ RequestedModel: "gpt-4.1",
+ ResolvedModel: "gpt-4.1",
+ Provider: "openai",
+ AliasUsed: false,
+ StatusCode: 500,
+ RequestID: "req-2",
+ ClientIP: "10.0.0.1",
+ Method: "POST",
+ Path: "/v1/responses",
+ Stream: false,
+ ErrorType: "server_error",
+ Data: nil,
},
})
normalized := strings.Join(strings.Fields(query), " ")
- wantQuery := "INSERT INTO audit_logs (id, timestamp, duration_ns, model, resolved_model, provider, alias_used, execution_plan_version_id, cache_type, status_code, request_id, auth_key_id, auth_method, client_ip, method, path, user_path, stream, error_type, data) VALUES ($1, $2, $3, $4, $5, $6, $7, $8, $9, $10, $11, $12, $13, $14, $15, $16, $17, $18, $19, $20), ($21, $22, $23, $24, $25, $26, $27, $28, $29, $30, $31, $32, $33, $34, $35, $36, $37, $38, $39, $40) ON CONFLICT (id) DO NOTHING"
+ wantQuery := "INSERT INTO audit_logs (id, timestamp, duration_ns, requested_model, resolved_model, provider, provider_name, alias_used, execution_plan_version_id, cache_type, status_code, request_id, auth_key_id, auth_method, client_ip, method, path, user_path, stream, error_type, data) VALUES ($1, $2, $3, $4, $5, $6, $7, $8, $9, $10, $11, $12, $13, $14, $15, $16, $17, $18, $19, $20, $21), ($22, $23, $24, $25, $26, $27, $28, $29, $30, $31, $32, $33, $34, $35, $36, $37, $38, $39, $40, $41, $42) ON CONFLICT (id) DO NOTHING"
if normalized != wantQuery {
t.Fatalf("query = %q, want %q", normalized, wantQuery)
}
- if got, want := len(args), 40; got != want {
+ if got, want := len(args), 42; got != want {
t.Fatalf("len(args) = %d, want %d", got, want)
}
if got := args[0]; got != "log-1" {
t.Fatalf("args[0] = %v, want log-1", got)
}
- if got := args[8]; got != CacheTypeExact {
- t.Fatalf("args[8] = %v, want %q", got, CacheTypeExact)
+ if got := args[6]; got != "primary-openai" {
+ t.Fatalf("args[6] = %v, want primary-openai", got)
}
- if got, ok := args[11].(string); !ok || got != "auth-key-1" {
- t.Fatalf("args[11] = (%T) %v, want (string) auth-key-1", args[11], args[11])
+ if got := args[9]; got != CacheTypeExact {
+ t.Fatalf("args[9] = %v, want %q", got, CacheTypeExact)
}
- if got, ok := args[12].(string); !ok || got != "" {
- t.Fatalf("args[12] = (%T) %v, want (string) \"\"", args[12], args[12])
+ if got, ok := args[12].(string); !ok || got != "auth-key-1" {
+ t.Fatalf("args[12] = (%T) %v, want (string) auth-key-1", args[12], args[12])
}
- if got, ok := args[15].(string); !ok || got != "/v1/chat/completions" {
- t.Fatalf("args[15] = (%T) %v, want (string) /v1/chat/completions", args[15], args[15])
+ if got, ok := args[13].(string); !ok || got != "" {
+ t.Fatalf("args[13] = (%T) %v, want (string) \"\"", args[13], args[13])
}
- if got, ok := args[16].(string); !ok || got != "/team/alpha" {
- t.Fatalf("args[16] = (%T) %v, want (string) /team/alpha", args[16], args[16])
+ if got, ok := args[16].(string); !ok || got != "/v1/chat/completions" {
+ t.Fatalf("args[16] = (%T) %v, want (string) /v1/chat/completions", args[16], args[16])
}
- if got := string(args[19].([]byte)); got != `{"user_agent":"test-agent"}` {
- t.Fatalf("args[19] = %q, want %q", got, `{"user_agent":"test-agent"}`)
+ if got, ok := args[17].(string); !ok || got != "/team/alpha" {
+ t.Fatalf("args[17] = (%T) %v, want (string) /team/alpha", args[17], args[17])
}
- if got := args[20]; got != "log-2" {
- t.Fatalf("args[20] = %v, want log-2", got)
+ if got := string(args[20].([]byte)); got != `{"user_agent":"test-agent"}` {
+ t.Fatalf("args[20] = %q, want %q", got, `{"user_agent":"test-agent"}`)
}
- if got, ok := args[31].(string); !ok || got != "" {
- t.Fatalf("args[31] = (%T) %v, want (string) \"\"", args[31], args[31])
+ if got := args[21]; got != "log-2" {
+ t.Fatalf("args[21] = %v, want log-2", got)
}
- if got, ok := args[32].(string); !ok || got != "" {
- t.Fatalf("args[32] = (%T) %v, want (string) \"\"", args[32], args[32])
+ if got, ok := args[33].(string); !ok || got != "" {
+ t.Fatalf("args[33] = (%T) %v, want (string) \"\"", args[33], args[33])
}
- if got := args[28]; got != nil {
- t.Fatalf("args[28] = %v, want nil cache type", got)
+ if got, ok := args[34].(string); !ok || got != "" {
+ t.Fatalf("args[34] = (%T) %v, want (string) \"\"", args[34], args[34])
}
- if got, ok := args[36].(string); !ok || got != "/" {
- t.Fatalf("args[36] = (%T) %v, want (string) \"/\"", args[36], args[36])
+ if got := args[30]; got != nil {
+ t.Fatalf("args[30] = %v, want nil cache type", got)
}
- dataJSON, ok := args[39].([]byte)
+ if got, ok := args[38].(string); !ok || got != "/" {
+ t.Fatalf("args[38] = (%T) %v, want (string) \"/\"", args[38], args[38])
+ }
+ dataJSON, ok := args[41].([]byte)
if !ok {
- t.Fatalf("args[39] has type %T, want []byte", args[39])
+ t.Fatalf("args[41] has type %T, want []byte", args[41])
}
if dataJSON != nil {
- t.Fatalf("args[39] = %v, want nil data", dataJSON)
+ t.Fatalf("args[41] = %v, want nil data", dataJSON)
}
}
diff --git a/internal/auditlog/store_sqlite.go b/internal/auditlog/store_sqlite.go
index 5ba4d2ae3..a1854f712 100644
--- a/internal/auditlog/store_sqlite.go
+++ b/internal/auditlog/store_sqlite.go
@@ -15,10 +15,12 @@ import (
// We chunk larger batches to avoid hitting this limit.
const (
maxSQLiteParams = 999
- columnsPerEntry = 20
+ columnsPerEntry = 21
maxEntriesPerBatch = maxSQLiteParams / columnsPerEntry // 49 entries
)
+const sqliteAuditLogTable = "audit_logs"
+
// SQLiteStore implements LogStore for SQLite databases.
type SQLiteStore struct {
db *sql.DB
@@ -41,9 +43,10 @@ func NewSQLiteStore(db *sql.DB, retentionDays int) (*SQLiteStore, error) {
id TEXT PRIMARY KEY,
timestamp DATETIME NOT NULL,
duration_ns INTEGER DEFAULT 0,
- model TEXT,
+ requested_model TEXT,
resolved_model TEXT,
provider TEXT,
+ provider_name TEXT,
alias_used INTEGER DEFAULT 0,
execution_plan_version_id TEXT,
cache_type TEXT,
@@ -64,8 +67,14 @@ func NewSQLiteStore(db *sql.DB, retentionDays int) (*SQLiteStore, error) {
return nil, fmt.Errorf("failed to create audit_logs table: %w", err)
}
+ if err := renameSQLiteAuditColumn(db, sqliteAuditLogTable, "model", "requested_model"); err != nil {
+ return nil, fmt.Errorf("failed to rename audit_logs.model to requested_model: %w", err)
+ }
+
migrations := []string{
+ "ALTER TABLE audit_logs ADD COLUMN requested_model TEXT",
"ALTER TABLE audit_logs ADD COLUMN resolved_model TEXT",
+ "ALTER TABLE audit_logs ADD COLUMN provider_name TEXT",
"ALTER TABLE audit_logs ADD COLUMN alias_used INTEGER DEFAULT 0",
"ALTER TABLE audit_logs ADD COLUMN execution_plan_version_id TEXT",
"ALTER TABLE audit_logs ADD COLUMN cache_type TEXT",
@@ -84,9 +93,11 @@ func NewSQLiteStore(db *sql.DB, retentionDays int) (*SQLiteStore, error) {
// Create indexes for common queries
indexes := []string{
"CREATE INDEX IF NOT EXISTS idx_audit_timestamp ON audit_logs(timestamp)",
- "CREATE INDEX IF NOT EXISTS idx_audit_model ON audit_logs(model)",
+ "DROP INDEX IF EXISTS idx_audit_model",
+ "CREATE INDEX IF NOT EXISTS idx_audit_requested_model ON audit_logs(requested_model)",
"CREATE INDEX IF NOT EXISTS idx_audit_status ON audit_logs(status_code)",
"CREATE INDEX IF NOT EXISTS idx_audit_provider ON audit_logs(provider)",
+ "CREATE INDEX IF NOT EXISTS idx_audit_provider_name ON audit_logs(provider_name)",
"CREATE INDEX IF NOT EXISTS idx_audit_execution_plan_version_id ON audit_logs(execution_plan_version_id)",
"CREATE INDEX IF NOT EXISTS idx_audit_request_id ON audit_logs(request_id)",
"CREATE INDEX IF NOT EXISTS idx_audit_auth_key_id ON audit_logs(auth_key_id)",
@@ -134,7 +145,7 @@ func (s *SQLiteStore) WriteBatch(ctx context.Context, entries []*LogEntry) error
values := make([]any, 0, len(chunk)*columnsPerEntry)
for j, e := range chunk {
- placeholders[j] = "(?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?)"
+ placeholders[j] = "(?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?)"
dataJSON := marshalLogData(e.Data, e.ID)
@@ -166,9 +177,10 @@ func (s *SQLiteStore) WriteBatch(ctx context.Context, entries []*LogEntry) error
e.ID,
e.Timestamp.UTC().Format(time.RFC3339Nano),
e.DurationNs,
- e.Model,
+ e.RequestedModel,
e.ResolvedModel,
e.Provider,
+ e.ProviderName,
aliasUsedInt,
e.ExecutionPlanVersionID,
cacheTypeValue,
@@ -186,7 +198,7 @@ func (s *SQLiteStore) WriteBatch(ctx context.Context, entries []*LogEntry) error
)
}
- query := `INSERT OR IGNORE INTO audit_logs (id, timestamp, duration_ns, model, resolved_model, provider, alias_used, execution_plan_version_id, cache_type, status_code,
+ query := `INSERT OR IGNORE INTO audit_logs (id, timestamp, duration_ns, requested_model, resolved_model, provider, provider_name, alias_used, execution_plan_version_id, cache_type, status_code,
request_id, auth_key_id, auth_method, client_ip, method, path, user_path, stream, error_type, data) VALUES ` +
strings.Join(placeholders, ",")
@@ -234,3 +246,45 @@ func (s *SQLiteStore) cleanup() {
slog.Info("cleaned up old audit logs", "deleted", rowsAffected)
}
}
+
+func renameSQLiteAuditColumn(db *sql.DB, tableName, from, to string) error {
+ if db == nil {
+ return nil
+ }
+ fromExists, err := sqliteColumnExists(db, tableName, from)
+ if err != nil || !fromExists {
+ return err
+ }
+ toExists, err := sqliteColumnExists(db, tableName, to)
+ if err != nil || toExists {
+ return err
+ }
+ _, err = db.Exec(fmt.Sprintf("ALTER TABLE %s RENAME COLUMN %s TO %s", tableName, from, to))
+ return err
+}
+
+func sqliteColumnExists(db *sql.DB, tableName, columnName string) (bool, error) {
+ rows, err := db.Query(fmt.Sprintf("PRAGMA table_info(%s)", tableName))
+ if err != nil {
+ return false, err
+ }
+ defer rows.Close()
+
+ for rows.Next() {
+ var (
+ cid int
+ name string
+ columnType string
+ notNull int
+ dfltValue any
+ pk int
+ )
+ if err := rows.Scan(&cid, &name, &columnType, ¬Null, &dfltValue, &pk); err != nil {
+ return false, err
+ }
+ if strings.EqualFold(name, columnName) {
+ return true, nil
+ }
+ }
+ return false, rows.Err()
+}
diff --git a/internal/auditlog/store_sqlite_test.go b/internal/auditlog/store_sqlite_test.go
index d61961ce8..1f2f4a906 100644
--- a/internal/auditlog/store_sqlite_test.go
+++ b/internal/auditlog/store_sqlite_test.go
@@ -35,17 +35,17 @@ func TestSQLiteStore_WriteBatch_NullDataPreservation(t *testing.T) {
// Create entries - one with nil Data, one with Data
entries := []*LogEntry{
{
- ID: "entry-nil-data",
- Timestamp: time.Now(),
- Model: "gpt-4",
- Provider: "openai",
- Data: nil, // This should become SQL NULL
+ ID: "entry-nil-data",
+ Timestamp: time.Now(),
+ RequestedModel: "gpt-4",
+ Provider: "openai",
+ Data: nil, // This should become SQL NULL
},
{
- ID: "entry-with-data",
- Timestamp: time.Now(),
- Model: "gpt-4",
- Provider: "openai",
+ ID: "entry-with-data",
+ Timestamp: time.Now(),
+ RequestedModel: "gpt-4",
+ Provider: "openai",
Data: &LogData{
UserAgent: "test-agent",
},
@@ -104,11 +104,11 @@ func TestSQLiteStore_WriteBatch_Chunking(t *testing.T) {
entries := make([]*LogEntry, numEntries)
for i := range numEntries {
entries[i] = &LogEntry{
- ID: fmt.Sprintf("entry-%03d", i),
- Timestamp: time.Now(),
- Model: "gpt-4",
- Provider: "openai",
- StatusCode: 200,
+ ID: fmt.Sprintf("entry-%03d", i),
+ Timestamp: time.Now(),
+ RequestedModel: "gpt-4",
+ Provider: "openai",
+ StatusCode: 200,
}
}
@@ -183,9 +183,9 @@ func TestSQLiteStore_WriteBatch_ExactBatchBoundary(t *testing.T) {
entries := make([]*LogEntry, numEntries)
for i := range numEntries {
entries[i] = &LogEntry{
- ID: fmt.Sprintf("exact-%03d", i),
- Timestamp: time.Now(),
- Model: "gpt-4",
+ ID: fmt.Sprintf("exact-%03d", i),
+ Timestamp: time.Now(),
+ RequestedModel: "gpt-4",
}
}
@@ -205,9 +205,9 @@ func TestSQLiteStore_WriteBatch_ExactBatchBoundary(t *testing.T) {
entries = make([]*LogEntry, maxEntriesPerBatch+1)
for i := 0; i <= maxEntriesPerBatch; i++ {
entries[i] = &LogEntry{
- ID: fmt.Sprintf("boundary-%03d", i),
- Timestamp: time.Now(),
- Model: "gpt-4",
+ ID: fmt.Sprintf("boundary-%03d", i),
+ Timestamp: time.Now(),
+ RequestedModel: "gpt-4",
}
}
@@ -236,13 +236,13 @@ func TestSQLiteStore_WriteBatch_PersistsAliasFields(t *testing.T) {
ctx := context.Background()
entry := &LogEntry{
- ID: "alias-entry",
- Timestamp: time.Now(),
- Model: "anthropic/claude-opus-4-6",
- ResolvedModel: "openai/gpt-5-nano",
- Provider: "openai",
- AliasUsed: true,
- StatusCode: 200,
+ ID: "alias-entry",
+ Timestamp: time.Now(),
+ RequestedModel: "anthropic/claude-opus-4-6",
+ ResolvedModel: "openai/gpt-5-nano",
+ Provider: "openai",
+ AliasUsed: true,
+ StatusCode: 200,
}
if err := store.WriteBatch(ctx, []*LogEntry{entry}); err != nil {
@@ -262,8 +262,8 @@ func TestSQLiteStore_WriteBatch_PersistsAliasFields(t *testing.T) {
t.Fatal("expected log entry, got nil")
return
}
- if logEntry.Model != entry.Model {
- t.Fatalf("Model = %q, want %q", logEntry.Model, entry.Model)
+ if logEntry.RequestedModel != entry.RequestedModel {
+ t.Fatalf("RequestedModel = %q, want %q", logEntry.RequestedModel, entry.RequestedModel)
}
if logEntry.ResolvedModel != entry.ResolvedModel {
t.Fatalf("ResolvedModel = %q, want %q", logEntry.ResolvedModel, entry.ResolvedModel)
@@ -292,7 +292,7 @@ func TestSQLiteReader_AllowsNullExecutionPlanVersionID(t *testing.T) {
now := time.Now().UTC().Format(time.RFC3339Nano)
if _, err := db.Exec(`
INSERT INTO audit_logs (
- id, timestamp, duration_ns, model, resolved_model, provider, alias_used, execution_plan_version_id,
+ id, timestamp, duration_ns, requested_model, resolved_model, provider, alias_used, execution_plan_version_id,
status_code, request_id, client_ip, method, path, stream, error_type, data
) VALUES (?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?)
`,
@@ -358,7 +358,7 @@ func TestSQLiteReader_GetLogsFiltersByUserPathSubtree(t *testing.T) {
now := time.Now().UTC().Format(time.RFC3339Nano)
_, err = db.Exec(`
INSERT INTO audit_logs (
- id, timestamp, duration_ns, model, resolved_model, provider, alias_used, execution_plan_version_id,
+ id, timestamp, duration_ns, requested_model, resolved_model, provider, alias_used, execution_plan_version_id,
status_code, request_id, client_ip, method, path, user_path, stream, error_type, data
) VALUES
(?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?),
@@ -436,7 +436,7 @@ func TestSQLiteReader_GetLogsRootUserPathIncludesLegacyNullRows(t *testing.T) {
now := time.Now().UTC().Format(time.RFC3339Nano)
_, err = db.Exec(`
INSERT INTO audit_logs (
- id, timestamp, duration_ns, model, resolved_model, provider, alias_used, execution_plan_version_id,
+ id, timestamp, duration_ns, requested_model, resolved_model, provider, alias_used, execution_plan_version_id,
status_code, request_id, client_ip, method, path, user_path, stream, error_type, data
) VALUES
(?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?),
@@ -509,17 +509,17 @@ func TestSQLiteStoreAndReader_PreserveCacheType(t *testing.T) {
now := time.Now()
if err := store.WriteBatch(ctx, []*LogEntry{
{
- ID: "cache-exact",
- Timestamp: now,
- Model: "gpt-4",
- Provider: "openai",
- CacheType: CacheTypeExact,
+ ID: "cache-exact",
+ Timestamp: now,
+ RequestedModel: "gpt-4",
+ Provider: "openai",
+ CacheType: CacheTypeExact,
},
{
- ID: "cache-none",
- Timestamp: now.Add(time.Second),
- Model: "gpt-4",
- Provider: "openai",
+ ID: "cache-none",
+ Timestamp: now.Add(time.Second),
+ RequestedModel: "gpt-4",
+ Provider: "openai",
},
}); err != nil {
t.Fatalf("WriteBatch failed: %v", err)
diff --git a/internal/auditlog/stream_wrapper.go b/internal/auditlog/stream_wrapper.go
index fe4e8567f..18848778c 100644
--- a/internal/auditlog/stream_wrapper.go
+++ b/internal/auditlog/stream_wrapper.go
@@ -84,7 +84,7 @@ func CreateStreamEntry(baseEntry *LogEntry) *LogEntry {
ID: baseEntry.ID,
Timestamp: baseEntry.Timestamp,
DurationNs: baseEntry.DurationNs,
- Model: baseEntry.Model,
+ RequestedModel: baseEntry.RequestedModel,
ResolvedModel: baseEntry.ResolvedModel,
Provider: baseEntry.Provider,
AliasUsed: baseEntry.AliasUsed,
diff --git a/internal/core/interfaces.go b/internal/core/interfaces.go
index 4bfdbdc84..d9072bdca 100644
--- a/internal/core/interfaces.go
+++ b/internal/core/interfaces.go
@@ -103,6 +103,12 @@ type RoutableProvider interface {
GetProviderType(model string) string
}
+// ProviderNameResolver is an optional interface for components that can map a
+// routed model selector back to the concrete configured provider instance name.
+type ProviderNameResolver interface {
+ GetProviderName(model string) string
+}
+
// AvailabilityChecker is an optional interface for providers that need
// to verify service availability before registration.
type AvailabilityChecker interface {
diff --git a/internal/core/request_model_resolution.go b/internal/core/request_model_resolution.go
index fd493c83f..4b9bd83fa 100644
--- a/internal/core/request_model_resolution.go
+++ b/internal/core/request_model_resolution.go
@@ -6,6 +6,7 @@ type RequestModelResolution struct {
Requested RequestedModelSelector
ResolvedSelector ModelSelector
ProviderType string
+ ProviderName string
AliasApplied bool
}
diff --git a/internal/providers/config.go b/internal/providers/config.go
index d5b083ec4..291c9f168 100644
--- a/internal/providers/config.go
+++ b/internal/providers/config.go
@@ -57,8 +57,9 @@ func applyProviderEnvVars(raw map[string]config.RawProviderConfig, discovery map
continue
}
- existing, exists := result[providerType]
- if exists {
+ targetKey, matched, ambiguous := findEnvOverlayTarget(result, providerType)
+ if matched {
+ existing := result[targetKey]
if apiKey != "" {
existing.APIKey = apiKey
}
@@ -70,7 +71,9 @@ func applyProviderEnvVars(raw map[string]config.RawProviderConfig, discovery map
if apiVersion != "" {
existing.APIVersion = apiVersion
}
- result[providerType] = existing
+ result[targetKey] = existing
+ } else if ambiguous {
+ continue
} else {
if spec.RequireBaseURL && explicitBaseURL == "" {
continue
@@ -87,6 +90,34 @@ func applyProviderEnvVars(raw map[string]config.RawProviderConfig, discovery map
return result
}
+func findEnvOverlayTarget(raw map[string]config.RawProviderConfig, providerType string) (string, bool, bool) {
+ if existing, ok := raw[providerType]; ok && rawProviderMatchesType(existing, providerType) {
+ return providerType, true, false
+ }
+
+ var matchedKey string
+ var matches int
+ for name, cfg := range raw {
+ if !rawProviderMatchesType(cfg, providerType) {
+ continue
+ }
+ matchedKey = name
+ matches++
+ if matches > 1 {
+ return "", false, true
+ }
+ }
+
+ if matches == 1 {
+ return matchedKey, true, false
+ }
+ return "", false, false
+}
+
+func rawProviderMatchesType(cfg config.RawProviderConfig, providerType string) bool {
+ return strings.TrimSpace(cfg.Type) == strings.TrimSpace(providerType)
+}
+
type providerEnvNames struct {
APIKey string
BaseURL string
diff --git a/internal/providers/config_test.go b/internal/providers/config_test.go
index 2174d8d29..2c18a954d 100644
--- a/internal/providers/config_test.go
+++ b/internal/providers/config_test.go
@@ -470,6 +470,49 @@ func TestApplyProviderEnvVars_EnvWinsOverYAML(t *testing.T) {
}
}
+func TestApplyProviderEnvVars_SingleCustomNamedProviderUsesTypeEnvVars(t *testing.T) {
+ t.Setenv("OPENAI_API_KEY", "sk-env-key")
+
+ raw := map[string]config.RawProviderConfig{
+ "openai_name": {Type: "openai"},
+ }
+ got := applyProviderEnvVars(raw, testDiscoveryConfigs)
+
+ provider, exists := got["openai_name"]
+ if !exists {
+ t.Fatal("expected custom-named openai provider to be preserved")
+ }
+ if provider.APIKey != "sk-env-key" {
+ t.Errorf("APIKey = %q, want sk-env-key", provider.APIKey)
+ }
+ if provider.BaseURL != testDiscoveryConfigs["openai"].DefaultBaseURL {
+ t.Errorf("BaseURL = %q, want %q", provider.BaseURL, testDiscoveryConfigs["openai"].DefaultBaseURL)
+ }
+ if _, exists := got["openai"]; exists {
+ t.Fatal("expected no duplicate auto-discovered openai provider")
+ }
+}
+
+func TestApplyProviderEnvVars_AmbiguousCustomNamedProvidersSkipTypeEnvOverlay(t *testing.T) {
+ t.Setenv("OPENAI_API_KEY", "sk-env-key")
+
+ raw := map[string]config.RawProviderConfig{
+ "openai-east": {Type: "openai", APIKey: "east-key", BaseURL: "https://east.example.com/v1"},
+ "openai-west": {Type: "openai", APIKey: "west-key", BaseURL: "https://west.example.com/v1"},
+ }
+ got := applyProviderEnvVars(raw, testDiscoveryConfigs)
+
+ if got["openai-east"].APIKey != "east-key" {
+ t.Errorf("openai-east APIKey = %q, want east-key", got["openai-east"].APIKey)
+ }
+ if got["openai-west"].APIKey != "west-key" {
+ t.Errorf("openai-west APIKey = %q, want west-key", got["openai-west"].APIKey)
+ }
+ if _, exists := got["openai"]; exists {
+ t.Fatal("expected no duplicate auto-discovered openai provider when multiple YAML providers share the type")
+ }
+}
+
func TestApplyProviderEnvVars_BaseURLEnvWinsOverYAML(t *testing.T) {
t.Setenv("OPENAI_BASE_URL", "https://env-override.com")
@@ -741,6 +784,33 @@ func TestResolveProviders_EmptyRaw_OnlyEnvVars(t *testing.T) {
}
}
+func TestResolveProviders_SingleCustomNamedProviderDoesNotDuplicateTypeKey(t *testing.T) {
+ t.Setenv("OPENAI_API_KEY", "sk-openai")
+
+ raw := map[string]config.RawProviderConfig{
+ "openai_name": {Type: "openai"},
+ }
+
+ got, filteredRaw := resolveProviders(raw, globalResilience, testDiscoveryConfigs)
+
+ provider, exists := got["openai_name"]
+ if !exists {
+ t.Fatal("expected openai_name provider in resolved providers")
+ }
+ if provider.APIKey != "sk-openai" {
+ t.Errorf("APIKey = %q, want sk-openai", provider.APIKey)
+ }
+ if provider.BaseURL != testDiscoveryConfigs["openai"].DefaultBaseURL {
+ t.Errorf("BaseURL = %q, want %q", provider.BaseURL, testDiscoveryConfigs["openai"].DefaultBaseURL)
+ }
+ if _, exists := got["openai"]; exists {
+ t.Fatal("expected no duplicate openai provider in resolved providers")
+ }
+ if _, exists := filteredRaw["openai"]; exists {
+ t.Fatal("expected no duplicate openai provider in filtered raw providers")
+ }
+}
+
func TestResolveProviders_NoProvidersNoEnvVars(t *testing.T) {
got, filteredRaw := resolveProviders(map[string]config.RawProviderConfig{}, globalResilience, testDiscoveryConfigs)
if len(got) != 0 {
diff --git a/internal/providers/registry.go b/internal/providers/registry.go
index 703ac7727..c6c1e50f5 100644
--- a/internal/providers/registry.go
+++ b/internal/providers/registry.go
@@ -636,6 +636,30 @@ func (r *ModelRegistry) GetProviderType(model string) string {
return ""
}
+// GetProviderName returns the concrete configured provider instance name for
+// the given model selector. Returns empty string if the model is not found.
+func (r *ModelRegistry) GetProviderName(model string) string {
+ r.mu.RLock()
+ defer r.mu.RUnlock()
+
+ providerName, modelID := splitModelSelector(model)
+ if providerName != "" {
+ if providerModels, ok := r.modelsByProvider[providerName]; ok {
+ if info, exists := providerModels[modelID]; exists {
+ return strings.TrimSpace(info.ProviderName)
+ }
+ }
+ if r.hasConfiguredProviderNameLocked(providerName) {
+ return ""
+ }
+ }
+
+ if info, ok := r.models[model]; ok {
+ return strings.TrimSpace(info.ProviderName)
+ }
+ return ""
+}
+
// ProviderByType returns the first registered provider for the given provider type.
// This lookup is independent of discovered models so provider-typed routes keep
// working even when a provider currently exposes zero models.
@@ -678,6 +702,22 @@ func (r *ModelRegistry) ProviderTypes() []string {
return result
}
+// ProviderNames returns the configured provider instance names in registration order.
+func (r *ModelRegistry) ProviderNames() []string {
+ r.mu.RLock()
+ defer r.mu.RUnlock()
+
+ result := make([]string, 0, len(r.providers))
+ for _, provider := range r.providers {
+ providerName := strings.TrimSpace(r.providerNames[provider])
+ if providerName == "" {
+ continue
+ }
+ result = append(result, providerName)
+ }
+ return result
+}
+
func splitModelSelector(model string) (providerName, modelID string) {
model = strings.TrimSpace(model)
if model == "" {
diff --git a/internal/providers/router.go b/internal/providers/router.go
index 20db2e95a..babf7390a 100644
--- a/internal/providers/router.go
+++ b/internal/providers/router.go
@@ -35,10 +35,18 @@ type providerTypeLister interface {
ProviderTypes() []string
}
+type providerNameLister interface {
+ ProviderNames() []string
+}
+
type publicModelLister interface {
ListPublicModels() []core.Model
}
+type modelWithProviderLister interface {
+ ListModelsWithProvider() []ModelWithProvider
+}
+
func registryUnavailableError(err error) error {
return core.NewProviderError("", http.StatusServiceUnavailable, err.Error(), err)
}
@@ -64,14 +72,139 @@ func (r *Router) checkReady() error {
return nil
}
-// resolveProvider validates readiness, parses the model selector, and finds the target provider.
-func (r *Router) resolveProvider(model, providerHint string) (core.Provider, core.ModelSelector, error) {
+// ResolveModel canonicalizes a requested selector into the concrete
+// provider-name-qualified selector used for execution.
+//
+// Resolution precedence is:
+// 1. configured provider name + model ID
+// 2. provider type + model ID
+// 3. raw slash-shaped model ID (only when provider was not explicit)
+// 4. default normalization fallback
+func (r *Router) ResolveModel(requested core.RequestedModelSelector) (core.ModelSelector, bool, error) {
if err := r.checkReady(); err != nil {
- return nil, core.ModelSelector{}, registryUnavailableError(err)
+ return core.ModelSelector{}, false, registryUnavailableError(err)
}
- selector, err := core.ParseModelSelector(model, providerHint)
+
+ requested = core.NewRequestedModelSelector(requested.Model, requested.ProviderHint)
+ selector, err := requested.Normalize()
if err != nil {
- return nil, core.ModelSelector{}, core.NewInvalidRequestError(err.Error(), err)
+ return core.ModelSelector{}, false, core.NewInvalidRequestError(err.Error(), err)
+ }
+
+ resolved := selector
+ if selector.Provider == "" {
+ if concrete, ok := r.resolveUnqualifiedSelector(selector); ok {
+ resolved = concrete
+ }
+ } else if concrete, ok := r.resolveQualifiedSelector(requested, selector); ok {
+ resolved = concrete
+ }
+
+ return resolved, resolved.QualifiedModel() != selector.QualifiedModel(), nil
+}
+
+func (r *Router) resolveUnqualifiedSelector(selector core.ModelSelector) (core.ModelSelector, bool) {
+ if selector.Provider != "" || strings.TrimSpace(selector.Model) == "" {
+ return core.ModelSelector{}, false
+ }
+
+ named, ok := r.lookup.(core.ProviderNameResolver)
+ if !ok {
+ return core.ModelSelector{}, false
+ }
+ providerName := strings.TrimSpace(named.GetProviderName(selector.Model))
+ if providerName == "" {
+ return core.ModelSelector{}, false
+ }
+ return core.ModelSelector{Provider: providerName, Model: selector.Model}, true
+}
+
+func (r *Router) resolveQualifiedSelector(requested core.RequestedModelSelector, selector core.ModelSelector) (core.ModelSelector, bool) {
+ models, ok := r.lookup.(modelWithProviderLister)
+ if !ok {
+ return core.ModelSelector{}, false
+ }
+
+ providerSegment := strings.TrimSpace(selector.Provider)
+ modelID := strings.TrimSpace(selector.Model)
+ if providerSegment == "" || modelID == "" {
+ return core.ModelSelector{}, false
+ }
+
+ entries := models.ListModelsWithProvider()
+
+ for _, entry := range entries {
+ if strings.TrimSpace(entry.ProviderName) != providerSegment {
+ continue
+ }
+ if strings.TrimSpace(entry.Model.ID) != modelID {
+ continue
+ }
+ return core.ModelSelector{Provider: entry.ProviderName, Model: entry.Model.ID}, true
+ }
+
+ for _, entry := range entries {
+ if strings.TrimSpace(entry.ProviderType) != providerSegment {
+ continue
+ }
+ if strings.TrimSpace(entry.Model.ID) != modelID {
+ continue
+ }
+ return core.ModelSelector{Provider: entry.ProviderName, Model: entry.Model.ID}, true
+ }
+
+ if requested.ExplicitProvider {
+ return core.ModelSelector{}, false
+ }
+ if r.hasConfiguredProviderName(providerSegment) {
+ return core.ModelSelector{}, false
+ }
+ if r.providerByTypeRegistry(providerSegment) != nil {
+ return core.ModelSelector{}, false
+ }
+
+ rawModelID := strings.TrimSpace(requested.Model)
+ if rawModelID == "" {
+ return core.ModelSelector{}, false
+ }
+ for _, entry := range entries {
+ if strings.TrimSpace(entry.Model.ID) != rawModelID {
+ continue
+ }
+ return core.ModelSelector{Provider: entry.ProviderName, Model: entry.Model.ID}, true
+ }
+
+ return core.ModelSelector{}, false
+}
+
+func (r *Router) hasConfiguredProviderName(providerName string) bool {
+ providerName = strings.TrimSpace(providerName)
+ if providerName == "" {
+ return false
+ }
+ if named, ok := r.lookup.(providerNameLister); ok {
+ for _, candidate := range named.ProviderNames() {
+ if strings.TrimSpace(candidate) == providerName {
+ return true
+ }
+ }
+ return false
+ }
+ if models, ok := r.lookup.(modelWithProviderLister); ok {
+ for _, entry := range models.ListModelsWithProvider() {
+ if strings.TrimSpace(entry.ProviderName) == providerName {
+ return true
+ }
+ }
+ }
+ return false
+}
+
+// resolveProvider validates readiness, parses the model selector, and finds the target provider.
+func (r *Router) resolveProvider(model, providerHint string) (core.Provider, core.ModelSelector, error) {
+ selector, _, err := r.ResolveModel(core.NewRequestedModelSelector(model, providerHint))
+ if err != nil {
+ return nil, core.ModelSelector{}, err
}
lookupModel := selector.QualifiedModel()
p := r.lookup.GetProvider(lookupModel)
@@ -259,10 +392,11 @@ func callEmbeddings(ctx context.Context, provider core.Provider, req *core.Embed
// Supports returns true if any provider supports the given model.
// Returns false if the lookup has no models loaded.
func (r *Router) Supports(model string) bool {
- if r.lookup.ModelCount() == 0 {
+ selector, _, err := r.ResolveModel(core.NewRequestedModelSelector(model, ""))
+ if err != nil {
return false
}
- return r.lookup.Supports(model)
+ return r.lookup.Supports(selector.QualifiedModel())
}
// ModelCount returns the number of models currently loaded into the router lookup.
@@ -374,7 +508,30 @@ func (r *Router) Embeddings(ctx context.Context, req *core.EmbeddingRequest) (*c
// GetProviderType returns the provider type string for the given model.
// Returns empty string if the model is not found.
func (r *Router) GetProviderType(model string) string {
- return r.lookup.GetProviderType(model)
+ selector, _, err := r.ResolveModel(core.NewRequestedModelSelector(model, ""))
+ if err != nil {
+ return ""
+ }
+ return r.lookup.GetProviderType(selector.QualifiedModel())
+}
+
+// GetProviderName returns the concrete configured provider instance name for
+// the given model selector. Returns empty string when unavailable.
+func (r *Router) GetProviderName(model string) string {
+ selector, _, err := r.ResolveModel(core.NewRequestedModelSelector(model, ""))
+ if err != nil {
+ return ""
+ }
+ if !r.lookup.Supports(selector.QualifiedModel()) {
+ return ""
+ }
+ if selector.Provider != "" {
+ return selector.Provider
+ }
+ if named, ok := r.lookup.(core.ProviderNameResolver); ok {
+ return named.GetProviderName(selector.QualifiedModel())
+ }
+ return ""
}
func (r *Router) providerByType(providerType string) core.Provider {
diff --git a/internal/providers/router_test.go b/internal/providers/router_test.go
index 8003bcdb6..4d63ff61f 100644
--- a/internal/providers/router_test.go
+++ b/internal/providers/router_test.go
@@ -31,6 +31,42 @@ func newMockLookup() *mockModelLookup {
}
}
+type registryModelEntry struct {
+ provider core.Provider
+ providerName string
+ providerType string
+ modelID string
+}
+
+func newTestRegistryWithModels(entries ...registryModelEntry) *ModelRegistry {
+ registry := NewModelRegistry()
+ for _, entry := range entries {
+ registry.RegisterProviderWithNameAndType(entry.provider, entry.providerName, entry.providerType)
+ }
+
+ registry.models = make(map[string]*ModelInfo)
+ registry.modelsByProvider = make(map[string]map[string]*ModelInfo)
+ for _, entry := range entries {
+ info := &ModelInfo{
+ Model: core.Model{
+ ID: entry.modelID,
+ Object: "model",
+ },
+ Provider: entry.provider,
+ ProviderName: entry.providerName,
+ ProviderType: entry.providerType,
+ }
+ if _, ok := registry.modelsByProvider[entry.providerName]; !ok {
+ registry.modelsByProvider[entry.providerName] = make(map[string]*ModelInfo)
+ }
+ registry.modelsByProvider[entry.providerName][entry.modelID] = info
+ if _, exists := registry.models[entry.modelID]; !exists {
+ registry.models[entry.modelID] = info
+ }
+ }
+ return registry
+}
+
func (m *mockModelLookup) addModel(model string, provider core.Provider, providerType string) {
m.models[model] = provider
m.providerTypes[model] = providerType
@@ -482,6 +518,91 @@ func TestRouterChatCompletion_PrefixedModelSelector(t *testing.T) {
}
}
+func TestRouterChatCompletion_PrefersProviderTypeSelectorOverRawSlashModel(t *testing.T) {
+ openAIResp := &core.ChatResponse{ID: "openai-test", Model: "gpt-5-nano"}
+ openRouterResp := &core.ChatResponse{ID: "openrouter", Model: "openai/gpt-5-nano"}
+ openAI := &mockProvider{name: "openai_test", chatResponse: openAIResp}
+ openRouter := &mockProvider{name: "openrouter", chatResponse: openRouterResp}
+
+ registry := newTestRegistryWithModels(
+ registryModelEntry{
+ provider: openAI,
+ providerName: "openai_test",
+ providerType: "openai",
+ modelID: "gpt-5-nano",
+ },
+ registryModelEntry{
+ provider: openRouter,
+ providerName: "openrouter",
+ providerType: "openrouter",
+ modelID: "openai/gpt-5-nano",
+ },
+ )
+
+ router, _ := NewRouter(registry)
+
+ resp, err := router.ChatCompletion(context.Background(), &core.ChatRequest{Model: "openai/gpt-5-nano"})
+ if err != nil {
+ t.Fatalf("unexpected error: %v", err)
+ }
+ if resp.ID != "openai-test" {
+ t.Fatalf("expected openai_test response, got %q", resp.ID)
+ }
+ if openAI.lastChatReq == nil || openAI.lastChatReq.Model != "gpt-5-nano" {
+ t.Fatalf("expected openai provider to receive raw model gpt-5-nano, got %#v", openAI.lastChatReq)
+ }
+ if openAI.lastChatReq.Provider != "" {
+ t.Fatalf("expected provider field to be stripped upstream, got %q", openAI.lastChatReq.Provider)
+ }
+ if openRouter.lastChatReq != nil {
+ t.Fatalf("expected openrouter provider to be bypassed, got %#v", openRouter.lastChatReq)
+ }
+ if got := router.GetProviderType("openai/gpt-5-nano"); got != "openai" {
+ t.Fatalf("GetProviderType() = %q, want %q", got, "openai")
+ }
+ if got := router.GetProviderName("openai/gpt-5-nano"); got != "openai_test" {
+ t.Fatalf("GetProviderName() = %q, want %q", got, "openai_test")
+ }
+}
+
+func TestRouterChatCompletion_ProviderQualifiedRawSlashModelStillWorks(t *testing.T) {
+ openAIResp := &core.ChatResponse{ID: "openai-test", Model: "gpt-5-nano"}
+ openRouterResp := &core.ChatResponse{ID: "openrouter", Model: "openai/gpt-5-nano"}
+ openAI := &mockProvider{name: "openai_test", chatResponse: openAIResp}
+ openRouter := &mockProvider{name: "openrouter", chatResponse: openRouterResp}
+
+ registry := newTestRegistryWithModels(
+ registryModelEntry{
+ provider: openAI,
+ providerName: "openai_test",
+ providerType: "openai",
+ modelID: "gpt-5-nano",
+ },
+ registryModelEntry{
+ provider: openRouter,
+ providerName: "openrouter",
+ providerType: "openrouter",
+ modelID: "openai/gpt-5-nano",
+ },
+ )
+
+ router, _ := NewRouter(registry)
+
+ resp, err := router.ChatCompletion(context.Background(), &core.ChatRequest{Model: "openrouter/openai/gpt-5-nano"})
+ if err != nil {
+ t.Fatalf("unexpected error: %v", err)
+ }
+ if resp.ID != "openrouter" {
+ t.Fatalf("expected openrouter response, got %q", resp.ID)
+ }
+ if openRouter.lastChatReq == nil || openRouter.lastChatReq.Model != "openai/gpt-5-nano" {
+ t.Fatalf("expected openrouter provider to receive raw slash model, got %#v", openRouter.lastChatReq)
+ }
+ if openRouter.lastChatReq.Provider != "" {
+ t.Fatalf("expected provider field to be stripped upstream, got %q", openRouter.lastChatReq.Provider)
+ }
+}
+
func TestRouterChatCompletion_ExplicitProviderKeepsSlashModelRaw(t *testing.T) {
groqResp := &core.ChatResponse{ID: "groq", Model: "openai/gpt-oss-120b"}
groq := &mockProvider{name: "groq", chatResponse: groqResp}
diff --git a/internal/responsecache/usage_hit.go b/internal/responsecache/usage_hit.go
index 4d22c6290..784e79547 100644
--- a/internal/responsecache/usage_hit.go
+++ b/internal/responsecache/usage_hit.go
@@ -28,10 +28,12 @@ func newUsageHitRecorder(logger usage.LoggerInterface, pricingResolver usage.Pri
model := ""
provider := ""
+ providerName := ""
if plan != nil {
provider = strings.TrimSpace(plan.ProviderType)
if plan.Resolution != nil {
model = strings.TrimSpace(plan.Resolution.ResolvedSelector.Model)
+ providerName = strings.TrimSpace(plan.Resolution.ProviderName)
}
}
if provider == "" {
@@ -54,6 +56,7 @@ func newUsageHitRecorder(logger usage.LoggerInterface, pricingResolver usage.Pri
if entry == nil {
return
}
+ entry.ProviderName = providerName
entry.UserPath = core.UserPathFromContext(ctx)
logger.Write(entry)
}
diff --git a/internal/server/execution_plan_helpers_test.go b/internal/server/execution_plan_helpers_test.go
index 073aab222..a413c4f81 100644
--- a/internal/server/execution_plan_helpers_test.go
+++ b/internal/server/execution_plan_helpers_test.go
@@ -57,7 +57,7 @@ func TestEnsureTranslatedRequestPlan_CompletesPartialPlanFromDecodedSelector(t *
assert.Equal(t, "gpt-4o-mini", storedPlan.Resolution.ResolvedSelector.Model)
}
}
- assert.Equal(t, "gpt-4o-mini", entry.Model)
+ assert.Equal(t, "gpt-4o-mini", entry.RequestedModel)
assert.Equal(t, "gpt-4o-mini", entry.ResolvedModel)
assert.Equal(t, "mock", entry.Provider)
}
diff --git a/internal/server/handlers_test.go b/internal/server/handlers_test.go
index 7497b96a7..88aabf6b7 100644
--- a/internal/server/handlers_test.go
+++ b/internal/server/handlers_test.go
@@ -306,6 +306,7 @@ type mockProvider struct {
streamData string
supportedModels []string
providerTypes map[string]string
+ providerNames map[string]string
batchCreateResponse *core.BatchResponse
batchCreateHints map[string]string
@@ -401,6 +402,25 @@ func (m *mockProvider) GetProviderType(model string) string {
return ""
}
+func (m *mockProvider) GetProviderName(model string) string {
+ selector, err := core.ParseModelSelector(model, "")
+ if err == nil && selector.Provider != "" {
+ if m.providerNames != nil {
+ if providerName, ok := m.providerNames[selector.QualifiedModel()]; ok {
+ return providerName
+ }
+ }
+ model = selector.Model
+ }
+
+ if m.providerNames != nil {
+ if providerName, ok := m.providerNames[model]; ok {
+ return providerName
+ }
+ }
+ return ""
+}
+
func (m *mockProvider) NativeFileProviderTypes() []string {
seen := make(map[string]struct{})
result := make([]string, 0, len(m.providerTypes))
@@ -2041,6 +2061,53 @@ func TestChatCompletionStreaming_FastPathUsesPassthroughForOpenAICompatibleProvi
}
}
+func TestChatCompletionStreaming_FastPathUsageCarriesResolvedProviderName(t *testing.T) {
+ streamData := "data: {\"id\":\"chatcmpl-123\",\"model\":\"gpt-4o-mini\",\"usage\":{\"prompt_tokens\":7,\"completion_tokens\":3,\"total_tokens\":10}}\n\ndata: [DONE]\n\n"
+ usageLog := &collectingUsageLogger{
+ config: usage.Config{Enabled: true},
+ }
+ mock := &mockProvider{
+ supportedModels: []string{"gpt-4o-mini"},
+ providerTypes: map[string]string{
+ "gpt-4o-mini": "openai",
+ },
+ providerNames: map[string]string{
+ "gpt-4o-mini": "openai_test",
+ },
+ passthroughResponse: &core.PassthroughResponse{
+ StatusCode: http.StatusOK,
+ Headers: map[string][]string{
+ "Content-Type": {"text/event-stream"},
+ },
+ Body: io.NopCloser(strings.NewReader(streamData)),
+ },
+ }
+
+ e := echo.New()
+ handler := NewHandler(mock, nil, usageLog, nil)
+
+ reqBody := `{"model":"gpt-4o-mini","stream":true,"messages":[{"role":"user","content":"Hi"}]}`
+ req := httptest.NewRequest(http.MethodPost, "/v1/chat/completions", strings.NewReader(reqBody))
+ req.Header.Set("Content-Type", "application/json")
+ rec := httptest.NewRecorder()
+ c := e.NewContext(req, rec)
+
+ err := handler.ChatCompletion(c)
+ if err != nil {
+ t.Fatalf("handler returned error: %v", err)
+ }
+
+ if rec.Code != http.StatusOK {
+ t.Fatalf("status = %d, want %d", rec.Code, http.StatusOK)
+ }
+ if len(usageLog.entries) != 1 {
+ t.Fatalf("usage entries = %d, want 1", len(usageLog.entries))
+ }
+ if got := usageLog.entries[0].ProviderName; got != "openai_test" {
+ t.Fatalf("ProviderName = %q, want openai_test", got)
+ }
+}
+
func TestChatCompletionStreaming_FastPathSkipsQualifiedModelRewrite(t *testing.T) {
streamData := "data: {\"id\":\"chatcmpl-123\",\"choices\":[{\"delta\":{\"content\":\"Hello\"}}]}\n\ndata: [DONE]\n\n"
provider := &capturingProvider{
@@ -2154,7 +2221,7 @@ func TestHandleStreamingResponse_FlushesEachChunk(t *testing.T) {
},
}
- err := handler.translatedInference().handleStreamingResponse(c, nil, "gpt-4o-mini", "openai", func() (io.ReadCloser, error) {
+ err := handler.translatedInference().handleStreamingResponse(c, nil, "gpt-4o-mini", "openai", "primary-openai", func() (io.ReadCloser, error) {
return stream, nil
})
if err != nil {
@@ -2246,7 +2313,7 @@ func TestHandleStreamingResponse_RecordsStreamingError(t *testing.T) {
Data: &auditlog.LogData{},
})
- err := handler.translatedInference().handleStreamingResponse(c, nil, "gpt-4o-mini", "openai", func() (io.ReadCloser, error) {
+ err := handler.translatedInference().handleStreamingResponse(c, nil, "gpt-4o-mini", "openai", "primary-openai", func() (io.ReadCloser, error) {
return &erroringReadCloser{
data: []byte("data: {\"id\":\"1\"}\n\n"),
err: expectedErr,
@@ -5550,8 +5617,8 @@ func TestProviderPassthrough_UsesPassthroughModelForAuditEntry(t *testing.T) {
if rec.Code != http.StatusOK {
t.Fatalf("status = %d, want 200", rec.Code)
}
- if entry.Model != "gpt-5-mini" {
- t.Fatalf("audit entry model = %q, want gpt-5-mini", entry.Model)
+ if entry.RequestedModel != "gpt-5-mini" {
+ t.Fatalf("audit entry requested model = %q, want gpt-5-mini", entry.RequestedModel)
}
if entry.Provider != "openai" {
t.Fatalf("audit entry provider = %q, want openai", entry.Provider)
diff --git a/internal/server/internal_chat_completion_executor.go b/internal/server/internal_chat_completion_executor.go
index 29f4a17a1..6f951992d 100644
--- a/internal/server/internal_chat_completion_executor.go
+++ b/internal/server/internal_chat_completion_executor.go
@@ -81,8 +81,10 @@ func (e *InternalChatCompletionExecutor) ChatCompletion(ctx context.Context, req
entry := e.newAuditEntry(ctx, requestID, requested)
var plan *core.ExecutionPlan
var cacheType string
+ var providerType string
+ var providerName string
defer func() {
- e.finishAuditEntry(ctx, entry, start, plan, req, resp, err, cacheType)
+ e.finishAuditEntry(ctx, entry, start, plan, req, resp, err, cacheType, providerType, providerName)
}()
resolution, err := resolveRequestModel(e.provider, e.modelResolver, requested)
@@ -102,13 +104,13 @@ func (e *InternalChatCompletionExecutor) ChatCompletion(ctx context.Context, req
ctx = e.service.withCacheRequestContext(ctx, plan)
execReq := cloneChatRequestForSelector(req, resolution.ResolvedSelector)
- resp, providerType, _, cacheType, err := e.executeChatCompletion(ctx, plan, execReq)
+ resp, providerType, providerName, _, cacheType, err = e.executeChatCompletion(ctx, plan, execReq)
if err != nil {
return nil, err
}
if cacheType == "" {
- e.service.logUsage(ctx, plan, resp.Model, providerType, func(pricing *core.ModelPricing) *usage.UsageEntry {
+ e.service.logUsage(ctx, plan, resp.Model, providerType, providerName, func(pricing *core.ModelPricing) *usage.UsageEntry {
return usage.ExtractFromChatResponse(resp, requestID, providerType, "/v1/chat/completions", pricing)
})
}
@@ -119,29 +121,30 @@ func (e *InternalChatCompletionExecutor) executeChatCompletion(
ctx context.Context,
plan *core.ExecutionPlan,
req *core.ChatRequest,
-) (*core.ChatResponse, string, bool, string, error) {
+) (*core.ChatResponse, string, string, bool, string, error) {
if e.service.responseCache == nil || (plan != nil && !plan.CacheEnabled()) {
- resp, providerType, usedFallback, err := e.service.executeChatCompletion(ctx, plan, req)
- return resp, providerType, usedFallback, "", err
+ resp, providerType, providerName, usedFallback, err := e.service.executeChatCompletion(ctx, plan, req)
+ return resp, providerType, providerName, usedFallback, "", err
}
body, err := marshalRequestBody(req)
if err != nil {
- resp, providerType, usedFallback, execErr := e.service.executeChatCompletion(ctx, plan, req)
+ resp, providerType, providerName, usedFallback, execErr := e.service.executeChatCompletion(ctx, plan, req)
if execErr != nil {
- return nil, "", false, "", execErr
+ return nil, "", "", false, "", execErr
}
- return resp, providerType, usedFallback, "", nil
+ return resp, providerType, providerName, usedFallback, "", nil
}
var (
resp *core.ChatResponse
providerType string
+ providerName string
usedFallback bool
)
result, err := e.service.responseCache.HandleInternalRequest(ctx, http.MethodPost, "/v1/chat/completions", body, func(c *echo.Context) error {
var execErr error
- resp, providerType, usedFallback, execErr = e.service.executeChatCompletion(c.Request().Context(), plan, req)
+ resp, providerType, providerName, usedFallback, execErr = e.service.executeChatCompletion(c.Request().Context(), plan, req)
if execErr != nil {
return execErr
}
@@ -151,16 +154,16 @@ func (e *InternalChatCompletionExecutor) executeChatCompletion(
return c.JSON(http.StatusOK, resp)
})
if err != nil {
- return nil, "", false, "", err
+ return nil, "", "", false, "", err
}
if result != nil && result.CacheType != "" {
var cached core.ChatResponse
if err := json.Unmarshal(result.Body, &cached); err != nil {
- return nil, "", false, "", err
+ return nil, "", "", false, "", err
}
- return &cached, plan.ProviderType, false, result.CacheType, nil
+ return &cached, plan.ProviderType, providerNameFromPlan(plan), false, result.CacheType, nil
}
- return resp, providerType, usedFallback, "", nil
+ return resp, providerType, providerName, usedFallback, "", nil
}
func (e *InternalChatCompletionExecutor) newAuditEntry(
@@ -187,7 +190,7 @@ func (e *InternalChatCompletionExecutor) newAuditEntry(
Data: &auditlog.LogData{},
}
if requestedModel := requested.RequestedQualifiedModel(); requestedModel != "" {
- entry.Model = requestedModel
+ entry.RequestedModel = requestedModel
}
return entry
}
@@ -201,6 +204,8 @@ func (e *InternalChatCompletionExecutor) finishAuditEntry(
resp *core.ChatResponse,
err error,
cacheType string,
+ providerType string,
+ providerName string,
) {
if entry == nil || e.logger == nil || !e.logger.Config().Enabled {
return
@@ -208,6 +213,7 @@ func (e *InternalChatCompletionExecutor) finishAuditEntry(
entry.DurationNs = time.Since(start).Nanoseconds()
auditlog.EnrichLogEntryWithExecutionPlan(entry, plan)
+ auditlog.EnrichLogEntryWithResolvedRoute(entry, qualifyExecutedModel(plan, chatResponseModel(resp), providerName), providerType, providerName)
auditlog.EnrichLogEntryWithRequestContext(entry, ctx)
if plan != nil && !plan.AuditEnabled() {
return
@@ -240,3 +246,10 @@ func (e *InternalChatCompletionExecutor) finishAuditEntry(
e.logger.Write(entry)
}
+
+func chatResponseModel(resp *core.ChatResponse) string {
+ if resp == nil {
+ return ""
+ }
+ return resp.Model
+}
diff --git a/internal/server/internal_chat_completion_executor_test.go b/internal/server/internal_chat_completion_executor_test.go
index 09784bf01..944ab9cb6 100644
--- a/internal/server/internal_chat_completion_executor_test.go
+++ b/internal/server/internal_chat_completion_executor_test.go
@@ -197,8 +197,8 @@ func TestInternalChatCompletionExecutor_DoesNotReuseParentExecutionPlanResolutio
t.Fatalf("audit entries = %d, want 1", len(logger.entries))
}
entry := logger.entries[0]
- if entry.Model != "openai/gpt-4o-mini" {
- t.Fatalf("audit requested model = %q, want openai/gpt-4o-mini", entry.Model)
+ if entry.RequestedModel != "openai/gpt-4o-mini" {
+ t.Fatalf("audit requested model = %q, want openai/gpt-4o-mini", entry.RequestedModel)
}
if entry.ResolvedModel != "openai/gpt-4o-mini" {
t.Fatalf("audit resolved model = %q, want openai/gpt-4o-mini", entry.ResolvedModel)
diff --git a/internal/server/model_validation_test.go b/internal/server/model_validation_test.go
index f65c9e23b..24e4e5529 100644
--- a/internal/server/model_validation_test.go
+++ b/internal/server/model_validation_test.go
@@ -705,7 +705,7 @@ func TestModelValidation_EnrichesAuditEntryWithRequestedModelOnResolutionError(t
assert.False(t, handlerCalled)
assert.Equal(t, http.StatusBadRequest, rec.Code)
assert.Contains(t, rec.Body.String(), "unsupported model: smart")
- assert.Equal(t, "smart", entry.Model)
+ assert.Equal(t, "smart", entry.RequestedModel)
assert.Equal(t, "", entry.ResolvedModel)
assert.Equal(t, "", entry.Provider)
assert.Equal(t, "invalid_request_error", entry.ErrorType)
diff --git a/internal/server/native_batch_support.go b/internal/server/native_batch_support.go
index 2f059bae3..a42a400df 100644
--- a/internal/server/native_batch_support.go
+++ b/internal/server/native_batch_support.go
@@ -59,22 +59,15 @@ func determineBatchExecutionSelection(provider core.RoutableProvider, resolver R
commonModel string
hasCommonModel = true
)
- resolver = effectiveRequestModelResolver(provider, resolver)
for i, item := range req.Requests {
requested, err := core.BatchItemRequestedModelSelector(req.Endpoint, item)
if err != nil {
return batchExecutionSelection{}, core.NewInvalidRequestError(fmt.Sprintf("batch item %d: %s", i, err.Error()), err)
}
- resolvedSelector, err := requested.Normalize()
+ resolvedSelector, _, err := resolveExecutionSelector(provider, resolver, requested)
if err != nil {
return batchExecutionSelection{}, core.NewInvalidRequestError(fmt.Sprintf("batch item %d: %s", i, err.Error()), err)
}
- if resolver != nil {
- resolvedSelector, _, err = resolver.ResolveModel(requested)
- if err != nil {
- return batchExecutionSelection{}, core.NewInvalidRequestError(fmt.Sprintf("batch item %d: %s", i, err.Error()), err)
- }
- }
model := resolvedSelector.QualifiedModel()
if model == "" {
return batchExecutionSelection{}, core.NewInvalidRequestError(fmt.Sprintf("batch item %d: model is required", i), nil)
diff --git a/internal/server/passthrough_service.go b/internal/server/passthrough_service.go
index 4a19a2d10..d151bb691 100644
--- a/internal/server/passthrough_service.go
+++ b/internal/server/passthrough_service.go
@@ -49,5 +49,5 @@ func (s *passthroughService) ProviderPassthrough(c *echo.Context) error {
} else {
auditlog.EnrichEntry(c, info.Model, providerType)
}
- return s.proxyPassthroughResponse(c, providerType, endpoint, info, resp)
+ return s.proxyPassthroughResponse(c, providerType, providerNameFromPlan(plan), endpoint, info, resp)
}
diff --git a/internal/server/passthrough_support.go b/internal/server/passthrough_support.go
index 3eab9dffa..e360d1e13 100644
--- a/internal/server/passthrough_support.go
+++ b/internal/server/passthrough_support.go
@@ -226,7 +226,7 @@ func passthroughAuditPath(c *echo.Context, providerType, endpoint string, info *
return passthroughStreamAuditPath("", providerType, endpoint)
}
-func (s *passthroughService) proxyPassthroughResponse(c *echo.Context, providerType, endpoint string, info *core.PassthroughRouteInfo, resp *core.PassthroughResponse) error {
+func (s *passthroughService) proxyPassthroughResponse(c *echo.Context, providerType, providerName, endpoint string, info *core.PassthroughRouteInfo, resp *core.PassthroughResponse) error {
if resp == nil || resp.Body == nil {
return handleError(c, core.NewProviderError(providerType, http.StatusBadGateway, "provider returned empty passthrough response", nil))
}
@@ -282,6 +282,7 @@ func (s *passthroughService) proxyPassthroughResponse(c *echo.Context, providerT
}
if s.usageLogger != nil && s.usageLogger.Config().Enabled && (plan == nil || plan.UsageEnabled()) {
if observer := usage.NewStreamUsageObserver(s.usageLogger, model, providerType, requestID, usagePath, s.pricingResolver, core.UserPathFromContext(c.Request().Context())); observer != nil {
+ observer.SetProviderName(providerName)
observers = append(observers, observer)
}
}
diff --git a/internal/server/request_model_resolution.go b/internal/server/request_model_resolution.go
index 6728a6f1d..5c8e55f36 100644
--- a/internal/server/request_model_resolution.go
+++ b/internal/server/request_model_resolution.go
@@ -21,33 +21,32 @@ type RequestFallbackResolver interface {
ResolveFallbacks(resolution *core.RequestModelResolution, op core.Operation) []core.ModelSelector
}
-func effectiveRequestModelResolver(provider core.RoutableProvider, resolver RequestModelResolver) RequestModelResolver {
- if resolver != nil {
- return resolver
+func resolvedProviderName(provider core.RoutableProvider, selector core.ModelSelector, fallback string) string {
+ fallback = strings.TrimSpace(fallback)
+ if provider == nil {
+ return fallback
}
- if providerResolver, ok := provider.(RequestModelResolver); ok {
- return providerResolver
+ if named, ok := provider.(core.ProviderNameResolver); ok {
+ if providerName := strings.TrimSpace(named.GetProviderName(selector.QualifiedModel())); providerName != "" {
+ return providerName
+ }
}
- return nil
+ return fallback
}
func resolveRequestModel(provider core.RoutableProvider, resolver RequestModelResolver, requested core.RequestedModelSelector) (*core.RequestModelResolution, error) {
requested = core.NewRequestedModelSelector(requested.Model, requested.ProviderHint)
- var (
- resolvedSelector core.ModelSelector
- aliasApplied bool
- err error
- )
-
- if effectiveResolver := effectiveRequestModelResolver(provider, resolver); effectiveResolver != nil {
- resolvedSelector, aliasApplied, err = effectiveResolver.ResolveModel(requested)
- } else {
- resolvedSelector, err = requested.Normalize()
- }
+ resolvedSelector, aliasApplied, err := resolveExecutionSelector(provider, resolver, requested)
if err != nil {
return nil, core.NewInvalidRequestError(err.Error(), err)
}
+ if resolvedSelector == (core.ModelSelector{}) {
+ resolvedSelector, err = requested.Normalize()
+ if err != nil {
+ return nil, core.NewInvalidRequestError(err.Error(), err)
+ }
+ }
resolvedModel := resolvedSelector.QualifiedModel()
if counted, ok := provider.(modelCountProvider); ok && counted.ModelCount() == 0 {
@@ -61,10 +60,49 @@ func resolveRequestModel(provider core.RoutableProvider, resolver RequestModelRe
Requested: requested,
ResolvedSelector: resolvedSelector,
ProviderType: strings.TrimSpace(provider.GetProviderType(resolvedModel)),
+ ProviderName: resolvedProviderName(provider, resolvedSelector, ""),
AliasApplied: aliasApplied,
}, nil
}
+func resolveExecutionSelector(
+ provider core.RoutableProvider,
+ resolver RequestModelResolver,
+ requested core.RequestedModelSelector,
+) (core.ModelSelector, bool, error) {
+ requested = core.NewRequestedModelSelector(requested.Model, requested.ProviderHint)
+
+ var (
+ resolvedSelector core.ModelSelector
+ aliasApplied bool
+ err error
+ )
+
+ if resolver != nil {
+ resolvedSelector, aliasApplied, err = resolver.ResolveModel(requested)
+ if err != nil {
+ return core.ModelSelector{}, false, err
+ }
+ requested = core.NewRequestedModelSelector(resolvedSelector.QualifiedModel(), "")
+ }
+
+ if providerResolver, ok := provider.(RequestModelResolver); ok {
+ var providerChanged bool
+ resolvedSelector, providerChanged, err = providerResolver.ResolveModel(requested)
+ if err != nil {
+ return core.ModelSelector{}, false, err
+ }
+ return resolvedSelector, aliasApplied || providerChanged, nil
+ }
+
+ if resolvedSelector != (core.ModelSelector{}) {
+ return resolvedSelector, aliasApplied, nil
+ }
+
+ resolvedSelector, err = requested.Normalize()
+ return resolvedSelector, aliasApplied, err
+}
+
func storeRequestModelResolution(c *echo.Context, resolution *core.RequestModelResolution) {
if c == nil || resolution == nil {
return
diff --git a/internal/server/request_model_resolution_test.go b/internal/server/request_model_resolution_test.go
new file mode 100644
index 000000000..60c2fbfdf
--- /dev/null
+++ b/internal/server/request_model_resolution_test.go
@@ -0,0 +1,156 @@
+package server
+
+import (
+ "context"
+ "io"
+ "testing"
+
+ "gomodel/internal/core"
+)
+
+type canonicalizingProvider struct {
+ resolved map[string]core.ModelSelector
+ types map[string]string
+ names map[string]string
+}
+
+func (p *canonicalizingProvider) ResolveModel(requested core.RequestedModelSelector) (core.ModelSelector, bool, error) {
+ key := requested.RequestedQualifiedModel()
+ if selector, ok := p.resolved[key]; ok {
+ return selector, selector.QualifiedModel() != key, nil
+ }
+ selector, err := requested.Normalize()
+ return selector, false, err
+}
+
+func (p *canonicalizingProvider) Supports(model string) bool {
+ _, ok := p.types[model]
+ return ok
+}
+
+func (p *canonicalizingProvider) GetProviderType(model string) string {
+ return p.types[model]
+}
+
+func (p *canonicalizingProvider) GetProviderName(model string) string {
+ return p.names[model]
+}
+
+func (p *canonicalizingProvider) ChatCompletion(_ context.Context, _ *core.ChatRequest) (*core.ChatResponse, error) {
+ return nil, nil
+}
+
+func (p *canonicalizingProvider) StreamChatCompletion(_ context.Context, _ *core.ChatRequest) (io.ReadCloser, error) {
+ return nil, nil
+}
+
+func (p *canonicalizingProvider) ListModels(_ context.Context) (*core.ModelsResponse, error) {
+ return nil, nil
+}
+
+func (p *canonicalizingProvider) Responses(_ context.Context, _ *core.ResponsesRequest) (*core.ResponsesResponse, error) {
+ return nil, nil
+}
+
+func (p *canonicalizingProvider) StreamResponses(_ context.Context, _ *core.ResponsesRequest) (io.ReadCloser, error) {
+ return nil, nil
+}
+
+func (p *canonicalizingProvider) Embeddings(_ context.Context, _ *core.EmbeddingRequest) (*core.EmbeddingResponse, error) {
+ return nil, nil
+}
+
+func TestResolveRequestModel_UsesResolvedProviderNameInsteadOfSelectorPrefix(t *testing.T) {
+ provider := &mockProvider{
+ supportedModels: []string{"gpt-5-nano"},
+ providerTypes: map[string]string{
+ "openai/gpt-5-nano": "openai",
+ },
+ providerNames: map[string]string{
+ "openai/gpt-5-nano": "openai_test",
+ },
+ }
+
+ resolution, err := resolveRequestModel(provider, nil, core.NewRequestedModelSelector("openai/gpt-5-nano", ""))
+ if err != nil {
+ t.Fatalf("resolveRequestModel() error = %v", err)
+ }
+
+ if got := resolution.ResolvedSelector.Provider; got != "openai" {
+ t.Fatalf("ResolvedSelector.Provider = %q, want %q", got, "openai")
+ }
+ if got := resolution.ProviderName; got != "openai_test" {
+ t.Fatalf("ProviderName = %q, want %q", got, "openai_test")
+ }
+}
+
+func TestResolveRequestModel_CanonicalizesProviderTypeSelectorToConcreteProviderName(t *testing.T) {
+ provider := &canonicalizingProvider{
+ resolved: map[string]core.ModelSelector{
+ "openai/gpt-5-nano": {Provider: "openai_test", Model: "gpt-5-nano"},
+ },
+ types: map[string]string{
+ "openai_test/gpt-5-nano": "openai",
+ },
+ names: map[string]string{
+ "openai_test/gpt-5-nano": "openai_test",
+ },
+ }
+
+ resolution, err := resolveRequestModel(provider, nil, core.NewRequestedModelSelector("openai/gpt-5-nano", ""))
+ if err != nil {
+ t.Fatalf("resolveRequestModel() error = %v", err)
+ }
+
+ if got := resolution.ResolvedQualifiedModel(); got != "openai_test/gpt-5-nano" {
+ t.Fatalf("ResolvedQualifiedModel = %q, want %q", got, "openai_test/gpt-5-nano")
+ }
+ if got := resolution.ProviderType; got != "openai" {
+ t.Fatalf("ProviderType = %q, want %q", got, "openai")
+ }
+ if got := resolution.ProviderName; got != "openai_test" {
+ t.Fatalf("ProviderName = %q, want %q", got, "openai_test")
+ }
+}
+
+type aliasResolverStub struct{}
+
+func (aliasResolverStub) ResolveModel(requested core.RequestedModelSelector) (core.ModelSelector, bool, error) {
+ if requested.RequestedQualifiedModel() == "anthropic/claude-opus-4-6" {
+ return core.ModelSelector{Provider: "openai", Model: "gpt-5-nano"}, true, nil
+ }
+ selector, err := requested.Normalize()
+ return selector, false, err
+}
+
+func TestResolveRequestModel_CanonicalizesAliasOutputThroughProviderResolver(t *testing.T) {
+ provider := &canonicalizingProvider{
+ resolved: map[string]core.ModelSelector{
+ "openai/gpt-5-nano": {Provider: "openai_test", Model: "gpt-5-nano"},
+ },
+ types: map[string]string{
+ "openai_test/gpt-5-nano": "openai",
+ },
+ names: map[string]string{
+ "openai_test/gpt-5-nano": "openai_test",
+ },
+ }
+
+ resolution, err := resolveRequestModel(provider, aliasResolverStub{}, core.NewRequestedModelSelector("anthropic/claude-opus-4-6", ""))
+ if err != nil {
+ t.Fatalf("resolveRequestModel() error = %v", err)
+ }
+
+ if !resolution.AliasApplied {
+ t.Fatal("AliasApplied = false, want true")
+ }
+ if got := resolution.ResolvedQualifiedModel(); got != "openai_test/gpt-5-nano" {
+ t.Fatalf("ResolvedQualifiedModel = %q, want %q", got, "openai_test/gpt-5-nano")
+ }
+ if got := resolution.ProviderType; got != "openai" {
+ t.Fatalf("ProviderType = %q, want %q", got, "openai")
+ }
+ if got := resolution.ProviderName; got != "openai_test" {
+ t.Fatalf("ProviderName = %q, want %q", got, "openai_test")
+ }
+}
diff --git a/internal/server/translated_inference_service.go b/internal/server/translated_inference_service.go
index 09631e2ac..d296a3479 100644
--- a/internal/server/translated_inference_service.go
+++ b/internal/server/translated_inference_service.go
@@ -79,7 +79,7 @@ func (s *translatedInferenceService) ChatCompletion(c *echo.Context) error {
func (s *translatedInferenceService) dispatchChatCompletion(c *echo.Context, req *core.ChatRequest, plan *core.ExecutionPlan) error {
ctx := c.Request().Context()
- streamReq, providerType, usageModel := s.resolveProviderAndModelFromPlan(c, plan, req.Model, req)
+ streamReq, providerType, providerName, usageModel := s.resolveProviderAndModelFromPlan(c, plan, req.Model, req)
requestID := requestIDFromContextOrHeader(c.Request())
if req.Stream {
@@ -88,25 +88,26 @@ func (s *translatedInferenceService) dispatchChatCompletion(c *echo.Context, req
return err
}
}
- stream, resolvedProviderType, resolvedUsageModel, usedFallback, err := s.streamChatCompletion(ctx, plan, streamReq, providerType, usageModel)
+ stream, resolvedProviderType, resolvedProviderName, resolvedUsageModel, usedFallback, err := s.streamChatCompletion(ctx, plan, streamReq, providerType, providerName, usageModel)
if err != nil {
return handleError(c, err)
}
if usedFallback {
markRequestFallbackUsed(c)
}
- return s.handleStreamingReadCloser(c, plan, resolvedUsageModel, resolvedProviderType, stream)
+ return s.handleStreamingReadCloser(c, plan, resolvedUsageModel, resolvedProviderType, resolvedProviderName, stream)
}
- resp, providerType, usedFallback, err := s.executeChatCompletion(ctx, plan, req)
+ resp, providerType, providerName, usedFallback, err := s.executeChatCompletion(ctx, plan, req)
if err != nil {
return handleError(c, err)
}
if usedFallback {
markRequestFallbackUsed(c)
}
+ auditlog.EnrichEntryWithResolvedRoute(c, qualifyExecutedModel(plan, resp.Model, providerName), providerType, providerName)
- s.logUsage(ctx, plan, resp.Model, providerType, func(pricing *core.ModelPricing) *usage.UsageEntry {
+ s.logUsage(ctx, plan, resp.Model, providerType, providerName, func(pricing *core.ModelPricing) *usage.UsageEntry {
return usage.ExtractFromChatResponse(resp, requestID, providerType, "/v1/chat/completions", pricing)
})
@@ -196,32 +197,33 @@ func (s *translatedInferenceService) withCacheRequestContext(ctx context.Context
func (s *translatedInferenceService) dispatchResponses(c *echo.Context, req *core.ResponsesRequest, plan *core.ExecutionPlan) error {
ctx := c.Request().Context()
- _, providerType, usageModel := s.resolveProviderAndModelFromPlan(c, plan, req.Model, nil)
+ _, providerType, providerName, usageModel := s.resolveProviderAndModelFromPlan(c, plan, req.Model, nil)
requestID := requestIDFromContextOrHeader(c.Request())
if req.Stream {
if (plan == nil || plan.UsageEnabled()) && s.shouldEnforceReturningUsageData() {
ctx = core.WithEnforceReturningUsageData(ctx, true)
}
- stream, resolvedProviderType, resolvedUsageModel, usedFallback, err := s.streamResponses(ctx, plan, req, providerType, usageModel)
+ stream, resolvedProviderType, resolvedProviderName, resolvedUsageModel, usedFallback, err := s.streamResponses(ctx, plan, req, providerType, providerName, usageModel)
if err != nil {
return handleError(c, err)
}
if usedFallback {
markRequestFallbackUsed(c)
}
- return s.handleStreamingReadCloser(c, plan, resolvedUsageModel, resolvedProviderType, stream)
+ return s.handleStreamingReadCloser(c, plan, resolvedUsageModel, resolvedProviderType, resolvedProviderName, stream)
}
- resp, providerType, usedFallback, err := s.executeResponses(ctx, plan, req)
+ resp, providerType, providerName, usedFallback, err := s.executeResponses(ctx, plan, req)
if err != nil {
return handleError(c, err)
}
if usedFallback {
markRequestFallbackUsed(c)
}
+ auditlog.EnrichEntryWithResolvedRoute(c, qualifyExecutedModel(plan, resp.Model, providerName), providerType, providerName)
- s.logUsage(ctx, plan, resp.Model, providerType, func(pricing *core.ModelPricing) *usage.UsageEntry {
+ s.logUsage(ctx, plan, resp.Model, providerType, providerName, func(pricing *core.ModelPricing) *usage.UsageEntry {
return usage.ExtractFromResponsesResponse(resp, requestID, providerType, "/v1/responses", pricing)
})
@@ -265,7 +267,7 @@ func (s *translatedInferenceService) tryFastPathStreamingChatPassthrough(c *echo
usageLogger: s.usageLogger,
pricingResolver: s.pricingResolver,
}
- return true, passthrough.proxyPassthroughResponse(c, providerType, endpoint, info, resp)
+ return true, passthrough.proxyPassthroughResponse(c, providerType, providerNameFromPlan(plan), endpoint, info, resp)
}
func (s *translatedInferenceService) canFastPathStreamingChatPassthrough(plan *core.ExecutionPlan, req *core.ChatRequest) bool {
@@ -335,12 +337,13 @@ func (s *translatedInferenceService) Embeddings(c *echo.Context) error {
ctx := c.Request().Context()
requestID := requestIDFromContextOrHeader(c.Request())
- resp, providerType, err := s.executeEmbeddings(ctx, plan, req)
+ resp, providerType, providerName, err := s.executeEmbeddings(ctx, plan, req)
if err != nil {
return handleError(c, err)
}
+ auditlog.EnrichEntryWithResolvedRoute(c, qualifyExecutedModel(plan, resp.Model, providerName), providerType, providerName)
- s.logUsage(ctx, plan, resp.Model, providerType, func(pricing *core.ModelPricing) *usage.UsageEntry {
+ s.logUsage(ctx, plan, resp.Model, providerType, providerName, func(pricing *core.ModelPricing) *usage.UsageEntry {
return usage.ExtractFromEmbeddingResponse(resp, requestID, providerType, "/v1/embeddings", pricing)
})
@@ -350,11 +353,12 @@ func (s *translatedInferenceService) Embeddings(c *echo.Context) error {
func (s *translatedInferenceService) handleStreamingReadCloser(
c *echo.Context,
plan *core.ExecutionPlan,
- model, provider string,
+ model, provider, providerName string,
stream io.ReadCloser,
) error {
auditlog.MarkEntryAsStreaming(c, true)
auditlog.EnrichEntryWithStream(c, true)
+ auditlog.EnrichEntryWithResolvedRoute(c, qualifyExecutedModel(plan, model, providerName), provider, providerName)
entry := auditlog.GetStreamEntryFromContext(c)
auditEnabled := s.logger != nil && s.logger.Config().Enabled && (plan == nil || plan.AuditEnabled())
@@ -373,7 +377,11 @@ func (s *translatedInferenceService) handleStreamingReadCloser(
observers = append(observers, auditlog.NewStreamLogObserver(s.logger, streamEntry, endpoint))
}
if s.usageLogger != nil && s.usageLogger.Config().Enabled && (plan == nil || plan.UsageEnabled()) {
- observers = append(observers, usage.NewStreamUsageObserver(s.usageLogger, model, provider, requestID, endpoint, s.pricingResolver, core.UserPathFromContext(c.Request().Context())))
+ usageObserver := usage.NewStreamUsageObserver(s.usageLogger, model, provider, requestID, endpoint, s.pricingResolver, core.UserPathFromContext(c.Request().Context()))
+ if usageObserver != nil {
+ usageObserver.SetProviderName(providerName)
+ observers = append(observers, usageObserver)
+ }
}
wrappedStream := streaming.NewObservedSSEStream(stream, observers...)
@@ -399,14 +407,14 @@ func (s *translatedInferenceService) handleStreamingReadCloser(
func (s *translatedInferenceService) handleStreamingResponse(
c *echo.Context,
plan *core.ExecutionPlan,
- model, provider string,
+ model, provider, providerName string,
streamFn func() (io.ReadCloser, error),
) error {
stream, err := streamFn()
if err != nil {
return handleError(c, err)
}
- return s.handleStreamingReadCloser(c, plan, model, provider, stream)
+ return s.handleStreamingReadCloser(c, plan, model, provider, providerName, stream)
}
//nolint:dupl // typed wrapper over the shared translated fallback executor
@@ -414,7 +422,7 @@ func (s *translatedInferenceService) executeChatCompletion(
ctx context.Context,
plan *core.ExecutionPlan,
req *core.ChatRequest,
-) (*core.ChatResponse, string, bool, error) {
+) (*core.ChatResponse, string, string, bool, error) {
return executeTranslatedWithFallback(ctx, s, plan, req, req.Model, req.Provider, cloneChatRequestForSelector,
func(ctx context.Context, req *core.ChatRequest) (*core.ChatResponse, string, error) {
resp, err := s.provider.ChatCompletion(ctx, req)
@@ -430,15 +438,15 @@ func (s *translatedInferenceService) streamChatCompletion(
ctx context.Context,
plan *core.ExecutionPlan,
req *core.ChatRequest,
- providerType, usageModel string,
-) (io.ReadCloser, string, string, bool, error) {
+ providerType, providerName, usageModel string,
+) (io.ReadCloser, string, string, string, bool, error) {
stream, err := s.provider.StreamChatCompletion(ctx, req)
if err == nil {
- return stream, providerType, usageModel, false, nil
+ return stream, providerType, providerName, usageModel, false, nil
}
- stream, resolvedProviderType, resolvedUsageModel, err := tryFallbackStream(ctx, s, plan, req.Model, req.Provider, err,
- func(selector core.ModelSelector, providerType string) (io.ReadCloser, string, string, error) {
+ stream, resolvedProviderType, resolvedProviderName, resolvedUsageModel, err := tryFallbackStream(ctx, s, plan, req.Model, req.Provider, err,
+ func(selector core.ModelSelector, providerType, providerName string) (io.ReadCloser, string, string, error) {
stream, err := s.provider.StreamChatCompletion(ctx, cloneChatRequestForSelector(req, selector))
if err != nil {
return nil, "", "", err
@@ -447,9 +455,9 @@ func (s *translatedInferenceService) streamChatCompletion(
},
)
if err != nil {
- return nil, "", "", false, err
+ return nil, "", "", "", false, err
}
- return stream, resolvedProviderType, resolvedUsageModel, true, nil
+ return stream, resolvedProviderType, resolvedProviderName, resolvedUsageModel, true, nil
}
//nolint:dupl // typed wrapper over the shared translated fallback executor
@@ -457,7 +465,7 @@ func (s *translatedInferenceService) executeResponses(
ctx context.Context,
plan *core.ExecutionPlan,
req *core.ResponsesRequest,
-) (*core.ResponsesResponse, string, bool, error) {
+) (*core.ResponsesResponse, string, string, bool, error) {
return executeTranslatedWithFallback(ctx, s, plan, req, req.Model, req.Provider, cloneResponsesRequestForSelector,
func(ctx context.Context, req *core.ResponsesRequest) (*core.ResponsesResponse, string, error) {
resp, err := s.provider.Responses(ctx, req)
@@ -473,15 +481,15 @@ func (s *translatedInferenceService) streamResponses(
ctx context.Context,
plan *core.ExecutionPlan,
req *core.ResponsesRequest,
- providerType, usageModel string,
-) (io.ReadCloser, string, string, bool, error) {
+ providerType, providerName, usageModel string,
+) (io.ReadCloser, string, string, string, bool, error) {
stream, err := s.provider.StreamResponses(ctx, req)
if err == nil {
- return stream, providerType, usageModel, false, nil
+ return stream, providerType, providerName, usageModel, false, nil
}
- stream, resolvedProviderType, resolvedUsageModel, err := tryFallbackStream(ctx, s, plan, req.Model, req.Provider, err,
- func(selector core.ModelSelector, providerType string) (io.ReadCloser, string, string, error) {
+ stream, resolvedProviderType, resolvedProviderName, resolvedUsageModel, err := tryFallbackStream(ctx, s, plan, req.Model, req.Provider, err,
+ func(selector core.ModelSelector, providerType, providerName string) (io.ReadCloser, string, string, error) {
stream, err := s.provider.StreamResponses(ctx, cloneResponsesRequestForSelector(req, selector))
if err != nil {
return nil, "", "", err
@@ -490,20 +498,21 @@ func (s *translatedInferenceService) streamResponses(
},
)
if err != nil {
- return nil, "", "", false, err
+ return nil, "", "", "", false, err
}
- return stream, resolvedProviderType, resolvedUsageModel, true, nil
+ return stream, resolvedProviderType, resolvedProviderName, resolvedUsageModel, true, nil
}
func (s *translatedInferenceService) executeEmbeddings(
ctx context.Context,
plan *core.ExecutionPlan,
req *core.EmbeddingRequest,
-) (*core.EmbeddingResponse, string, error) {
+) (*core.EmbeddingResponse, string, string, error) {
providerType := providerTypeFromPlan(plan)
+ providerName := providerNameFromPlan(plan)
resp, err := s.provider.Embeddings(ctx, req)
if err == nil {
- return resp, responseProviderType(providerType, resp.Provider), nil
+ return resp, responseProviderType(providerType, resp.Provider), providerName, nil
}
return s.tryFallbackEmbeddings(ctx, plan, req, err)
@@ -514,16 +523,16 @@ func (s *translatedInferenceService) tryFallbackEmbeddings(
plan *core.ExecutionPlan,
req *core.EmbeddingRequest,
primaryErr error,
-) (*core.EmbeddingResponse, string, error) {
+) (*core.EmbeddingResponse, string, string, error) {
// Embeddings fallback is intentionally disabled until the shared model
// contract can prove vector-size compatibility for alternates.
- return nil, "", primaryErr
+ return nil, "", "", primaryErr
}
func (s *translatedInferenceService) logUsage(
ctx context.Context,
plan *core.ExecutionPlan,
- model, providerType string,
+ model, providerType, providerName string,
extractFn func(*core.ModelPricing) *usage.UsageEntry,
) {
if s.usageLogger == nil || !s.usageLogger.Config().Enabled || (plan != nil && !plan.UsageEnabled()) {
@@ -534,6 +543,7 @@ func (s *translatedInferenceService) logUsage(
pricing = s.pricingResolver.ResolvePricing(model, providerType)
}
if entry := extractFn(pricing); entry != nil {
+ entry.ProviderName = strings.TrimSpace(providerName)
entry.UserPath = core.UserPathFromContext(ctx)
s.usageLogger.Write(entry)
}
@@ -552,15 +562,18 @@ func (s *translatedInferenceService) fallbackSelectors(plan *core.ExecutionPlan)
func (s *translatedInferenceService) providerTypeForSelector(selector core.ModelSelector, fallback string) string {
fallback = strings.TrimSpace(fallback)
- if provider := strings.TrimSpace(selector.Provider); provider != "" {
- return provider
- }
if s.provider == nil {
+ if provider := strings.TrimSpace(selector.Provider); provider != "" {
+ return provider
+ }
return fallback
}
if providerType := strings.TrimSpace(s.provider.GetProviderType(selector.QualifiedModel())); providerType != "" {
return providerType
}
+ if provider := strings.TrimSpace(selector.Provider); provider != "" {
+ return provider
+ }
return fallback
}
@@ -569,8 +582,9 @@ func (s *translatedInferenceService) resolveProviderAndModelFromPlan(
plan *core.ExecutionPlan,
fallbackModel string,
req *core.ChatRequest,
-) (*core.ChatRequest, string, string) {
+) (*core.ChatRequest, string, string, string) {
providerType := GetProviderType(c)
+ providerName := providerNameFromPlan(plan)
if plan != nil {
if plannedProviderType := strings.TrimSpace(plan.ProviderType); plannedProviderType != "" {
providerType = plannedProviderType
@@ -579,7 +593,7 @@ func (s *translatedInferenceService) resolveProviderAndModelFromPlan(
model := resolvedModelFromPlan(plan, fallbackModel)
if req == nil || !req.Stream || (plan != nil && !plan.UsageEnabled()) || !s.shouldEnforceReturningUsageData() {
- return req, providerType, model
+ return req, providerType, providerName, model
}
streamReq := cloneChatRequestForStreamUsage(req)
@@ -587,7 +601,7 @@ func (s *translatedInferenceService) resolveProviderAndModelFromPlan(
streamReq.StreamOptions = &core.StreamOptions{}
}
streamReq.StreamOptions.IncludeUsage = true
- return streamReq, providerType, model
+ return streamReq, providerType, providerName, model
}
func recordStreamingError(streamEntry *auditlog.LogEntry, model, provider, path, requestID string, err error) {
@@ -648,6 +662,42 @@ func cloneResponsesRequestForSelector(req *core.ResponsesRequest, selector core.
return &cloned
}
+func providerNameFromPlan(plan *core.ExecutionPlan) string {
+ if plan == nil || plan.Resolution == nil {
+ return ""
+ }
+ return strings.TrimSpace(plan.Resolution.ProviderName)
+}
+
+func resolvedModelPrefix(plan *core.ExecutionPlan, providerName string) string {
+ if providerName = strings.TrimSpace(providerName); providerName != "" {
+ return providerName
+ }
+ if plan == nil || plan.Resolution == nil {
+ return ""
+ }
+ if providerName = strings.TrimSpace(plan.Resolution.ProviderName); providerName != "" {
+ return providerName
+ }
+ return strings.TrimSpace(plan.Resolution.ResolvedSelector.Provider)
+}
+
+func qualifyModelWithProvider(model, providerName string) string {
+ model = strings.TrimSpace(model)
+ providerName = strings.TrimSpace(providerName)
+ if model == "" {
+ return ""
+ }
+ if providerName == "" || strings.HasPrefix(model, providerName+"/") {
+ return model
+ }
+ return providerName + "/" + model
+}
+
+func qualifyExecutedModel(plan *core.ExecutionPlan, model, providerName string) string {
+ return qualifyModelWithProvider(model, resolvedModelPrefix(plan, providerName))
+}
+
func markRequestFallbackUsed(c *echo.Context) {
if c == nil || c.Request() == nil {
return
@@ -706,13 +756,13 @@ func tryFallbackResponse[T any](
plan *core.ExecutionPlan,
model, provider string,
primaryErr error,
- call func(selector core.ModelSelector, providerType string) (T, string, error),
-) (T, string, bool, error) {
+ call func(selector core.ModelSelector, providerType, providerName string) (T, string, error),
+) (T, string, string, bool, error) {
var zero T
fallbacks := s.fallbackSelectors(plan)
if len(fallbacks) == 0 || !shouldAttemptFallback(primaryErr) {
- return zero, "", false, primaryErr
+ return zero, "", "", false, primaryErr
}
requestID := strings.TrimSpace(core.GetRequestID(ctx))
@@ -721,6 +771,7 @@ func tryFallbackResponse[T any](
for _, selector := range fallbacks {
qualified := selector.QualifiedModel()
providerType := s.providerTypeForSelector(selector, providerTypeFromPlan(plan))
+ providerName := resolvedProviderName(s.provider, selector, providerNameFromPlan(plan))
slog.Warn("primary model attempt failed, trying fallback",
"request_id", requestID,
"from", primaryModel,
@@ -729,7 +780,7 @@ func tryFallbackResponse[T any](
"error", lastErr,
)
- resp, resolvedProviderType, err := call(selector, providerType)
+ resp, resolvedProviderType, err := call(selector, providerType, providerName)
if err == nil {
slog.Info("fallback model attempt succeeded",
"request_id", requestID,
@@ -737,12 +788,12 @@ func tryFallbackResponse[T any](
"to", qualified,
"provider_type", resolvedProviderType,
)
- return resp, resolvedProviderType, true, nil
+ return resp, resolvedProviderType, providerName, true, nil
}
lastErr = err
}
- return zero, "", false, lastErr
+ return zero, "", "", false, lastErr
}
func executeWithFallbackResponse[T any](
@@ -750,12 +801,12 @@ func executeWithFallbackResponse[T any](
s *translatedInferenceService,
plan *core.ExecutionPlan,
model, provider string,
- primary func() (T, string, error),
- fallback func(selector core.ModelSelector, providerType string) (T, string, error),
-) (T, string, bool, error) {
- resp, resolvedProviderType, err := primary()
+ primary func() (T, string, string, error),
+ fallback func(selector core.ModelSelector, providerType, providerName string) (T, string, error),
+) (T, string, string, bool, error) {
+ resp, resolvedProviderType, resolvedProviderName, err := primary()
if err == nil {
- return resp, resolvedProviderType, false, nil
+ return resp, resolvedProviderType, resolvedProviderName, false, nil
}
return tryFallbackResponse(ctx, s, plan, model, provider, err, fallback)
}
@@ -768,17 +819,17 @@ func executeTranslatedWithFallback[Req any, Resp any](
model, provider string,
cloneForSelector func(Req, core.ModelSelector) Req,
call func(context.Context, Req) (Resp, string, error),
-) (Resp, string, bool, error) {
+) (Resp, string, string, bool, error) {
return executeWithFallbackResponse(ctx, s, plan, model, provider,
- func() (Resp, string, error) {
+ func() (Resp, string, string, error) {
resp, responseProvider, err := call(ctx, req)
if err != nil {
var zero Resp
- return zero, "", err
+ return zero, "", "", err
}
- return resp, responseProviderType(providerTypeFromPlan(plan), responseProvider), nil
+ return resp, responseProviderType(providerTypeFromPlan(plan), responseProvider), providerNameFromPlan(plan), nil
},
- func(selector core.ModelSelector, providerType string) (Resp, string, error) {
+ func(selector core.ModelSelector, providerType, providerName string) (Resp, string, error) {
resp, responseProvider, err := call(ctx, cloneForSelector(req, selector))
if err != nil {
var zero Resp
@@ -795,11 +846,11 @@ func tryFallbackStream(
plan *core.ExecutionPlan,
model, provider string,
primaryErr error,
- call func(selector core.ModelSelector, providerType string) (io.ReadCloser, string, string, error),
-) (io.ReadCloser, string, string, error) {
+ call func(selector core.ModelSelector, providerType, providerName string) (io.ReadCloser, string, string, error),
+) (io.ReadCloser, string, string, string, error) {
fallbacks := s.fallbackSelectors(plan)
if len(fallbacks) == 0 || !shouldAttemptFallback(primaryErr) {
- return nil, "", "", primaryErr
+ return nil, "", "", "", primaryErr
}
requestID := strings.TrimSpace(core.GetRequestID(ctx))
@@ -808,6 +859,7 @@ func tryFallbackStream(
for _, selector := range fallbacks {
qualified := selector.QualifiedModel()
providerType := s.providerTypeForSelector(selector, providerTypeFromPlan(plan))
+ providerName := resolvedProviderName(s.provider, selector, providerNameFromPlan(plan))
slog.Warn("primary model attempt failed, trying fallback stream",
"request_id", requestID,
"from", primaryModel,
@@ -816,7 +868,7 @@ func tryFallbackStream(
"error", lastErr,
)
- stream, resolvedProviderType, usageModel, err := call(selector, providerType)
+ stream, resolvedProviderType, usageModel, err := call(selector, providerType, providerName)
if err == nil {
slog.Info("fallback stream attempt succeeded",
"request_id", requestID,
@@ -824,12 +876,12 @@ func tryFallbackStream(
"to", qualified,
"provider_type", resolvedProviderType,
)
- return stream, resolvedProviderType, usageModel, nil
+ return stream, resolvedProviderType, providerName, usageModel, nil
}
lastErr = err
}
- return nil, "", "", lastErr
+ return nil, "", "", "", lastErr
}
func shouldAttemptFallback(err error) bool {
diff --git a/internal/server/translated_inference_service_test.go b/internal/server/translated_inference_service_test.go
index eef65c89c..14a12b07c 100644
--- a/internal/server/translated_inference_service_test.go
+++ b/internal/server/translated_inference_service_test.go
@@ -38,7 +38,7 @@ func TestTranslatedInferenceService_LogUsageSkipsWhenExecutionPlanDisablesUsage(
Guardrails: true,
},
},
- }, "gpt-5-nano", "openai", func(*core.ModelPricing) *usage.UsageEntry {
+ }, "gpt-5-nano", "openai", "primary-openai", func(*core.ModelPricing) *usage.UsageEntry {
return &usage.UsageEntry{ID: "usage-1"}
})
@@ -58,6 +58,39 @@ func TestTranslatedInferenceService_ProviderTypeForSelectorPrefersExplicitProvid
}
}
+func TestTranslatedInferenceService_ProviderTypeForSelectorCanonicalizesProviderNameSelectors(t *testing.T) {
+ service := &translatedInferenceService{
+ provider: &mockProvider{
+ supportedModels: []string{"gpt-4o"},
+ providerTypes: map[string]string{
+ "openai_test/gpt-4o": "openai",
+ },
+ providerNames: map[string]string{
+ "openai_test/gpt-4o": "openai_test",
+ },
+ },
+ }
+
+ got := service.providerTypeForSelector(core.ModelSelector{Provider: "openai_test", Model: "gpt-4o"}, "anthropic")
+ if got != "openai" {
+ t.Fatalf("providerTypeForSelector() = %q, want %q", got, "openai")
+ }
+}
+
+func TestQualifyModelWithProvider_PrefixesSlashModelIDs(t *testing.T) {
+ got := qualifyModelWithProvider("openai/gpt-4o-mini", "openrouter")
+ if got != "openrouter/openai/gpt-4o-mini" {
+ t.Fatalf("qualifyModelWithProvider() = %q, want %q", got, "openrouter/openai/gpt-4o-mini")
+ }
+}
+
+func TestQualifyModelWithProvider_KeepsAlreadyQualifiedModelIDs(t *testing.T) {
+ got := qualifyModelWithProvider("openrouter/openai/gpt-4o-mini", "openrouter")
+ if got != "openrouter/openai/gpt-4o-mini" {
+ t.Fatalf("qualifyModelWithProvider() = %q, want unchanged model", got)
+ }
+}
+
func TestTranslatedInferenceService_LogUsageAssignsUserPathFromContext(t *testing.T) {
logger := &usageCaptureLogger{
config: usage.Config{Enabled: true},
@@ -70,7 +103,7 @@ func TestTranslatedInferenceService_LogUsageAssignsUserPathFromContext(t *testin
UserPath: "/team/alpha",
})
- service.logUsage(ctx, nil, "gpt-5-nano", "openai", func(*core.ModelPricing) *usage.UsageEntry {
+ service.logUsage(ctx, nil, "gpt-5-nano", "openai", "primary-openai", func(*core.ModelPricing) *usage.UsageEntry {
return &usage.UsageEntry{ID: "usage-1"}
})
@@ -80,6 +113,9 @@ func TestTranslatedInferenceService_LogUsageAssignsUserPathFromContext(t *testin
if got := logger.entries[0].UserPath; got != "/team/alpha" {
t.Fatalf("UserPath = %q, want /team/alpha", got)
}
+ if got := logger.entries[0].ProviderName; got != "primary-openai" {
+ t.Fatalf("ProviderName = %q, want primary-openai", got)
+ }
}
func TestTranslatedInferenceService_WithCacheRequestContextClearsInheritedGuardrailsHash(t *testing.T) {
diff --git a/internal/usage/cache_type.go b/internal/usage/cache_type.go
index 1e36ba690..2f338c00c 100644
--- a/internal/usage/cache_type.go
+++ b/internal/usage/cache_type.go
@@ -46,11 +46,13 @@ func normalizedUsageEntryForStorage(entry *UsageEntry) *UsageEntry {
}
normalized := normalizeCacheType(entry.CacheType)
- if normalized == entry.CacheType {
+ providerName := strings.TrimSpace(entry.ProviderName)
+ if normalized == entry.CacheType && providerName == entry.ProviderName {
return entry
}
cloned := *entry
cloned.CacheType = normalized
+ cloned.ProviderName = providerName
return &cloned
}
diff --git a/internal/usage/reader.go b/internal/usage/reader.go
index 0d42a8248..e2461e68c 100644
--- a/internal/usage/reader.go
+++ b/internal/usage/reader.go
@@ -2,6 +2,7 @@ package usage
import (
"context"
+ "strings"
"time"
)
@@ -30,6 +31,7 @@ type UsageSummary struct {
type ModelUsage struct {
Model string `json:"model"`
Provider string `json:"provider"`
+ ProviderName string `json:"provider_name,omitempty"`
InputTokens int64 `json:"input_tokens"`
OutputTokens int64 `json:"output_tokens"`
InputCost *float64 `json:"input_cost"`
@@ -55,7 +57,7 @@ type DailyUsage struct {
type UsageLogParams struct {
UsageQueryParams // embed date range
Model string // filter by model (optional)
- Provider string // filter by provider (optional)
+ Provider string // filter by provider name or provider type (optional)
Search string // free-text search on model/provider/request_id
Limit int // page size (default 50, max 200)
Offset int // pagination offset
@@ -69,6 +71,7 @@ type UsageLogEntry struct {
Timestamp time.Time `json:"timestamp"`
Model string `json:"model"`
Provider string `json:"provider"`
+ ProviderName string `json:"provider_name,omitempty"`
Endpoint string `json:"endpoint"`
UserPath string `json:"user_path,omitempty"`
CacheType string `json:"cache_type,omitempty"`
@@ -138,3 +141,10 @@ type UsageReader interface {
// GetCacheOverview returns cached-only aggregates for the admin dashboard.
GetCacheOverview(ctx context.Context, params UsageQueryParams) (*CacheOverview, error)
}
+
+func displayUsageProviderName(providerName, provider string) string {
+ if trimmed := strings.TrimSpace(providerName); trimmed != "" {
+ return trimmed
+ }
+ return strings.TrimSpace(provider)
+}
diff --git a/internal/usage/reader_helpers.go b/internal/usage/reader_helpers.go
index c1d113d97..995b6e6e4 100644
--- a/internal/usage/reader_helpers.go
+++ b/internal/usage/reader_helpers.go
@@ -22,6 +22,12 @@ func buildWhereClause(conditions []string) string {
return " WHERE " + strings.Join(conditions, " AND ")
}
+// usageGroupedProviderNameSQL returns a SQL expression that collapses blank
+// provider_name values to the canonical provider before grouping.
+func usageGroupedProviderNameSQL(providerNameColumn, providerColumn string) string {
+ return "COALESCE(NULLIF(TRIM(" + providerNameColumn + "), ''), " + providerColumn + ")"
+}
+
// clampLimitOffset normalises pagination parameters:
// - limit defaults to 50 and is capped at 200
// - offset floors at 0
diff --git a/internal/usage/reader_mongodb.go b/internal/usage/reader_mongodb.go
index f2a235de6..3f94b3036 100644
--- a/internal/usage/reader_mongodb.go
+++ b/internal/usage/reader_mongodb.go
@@ -96,10 +96,12 @@ func (r *MongoDBReader) GetUsageByModel(ctx context.Context, params UsageQueryPa
pipeline = append(pipeline, bson.D{{Key: "$match", Value: matchFilters}})
}
+ providerNameExpr := mongoUsageGroupedProviderNameExpr()
pipeline = append(pipeline, bson.D{{Key: "$group", Value: bson.D{
{Key: "_id", Value: bson.D{
{Key: "model", Value: "$model"},
{Key: "provider", Value: "$provider"},
+ {Key: "provider_name", Value: providerNameExpr},
}},
{Key: "input_tokens", Value: bson.D{{Key: "$sum", Value: "$input_tokens"}}},
{Key: "output_tokens", Value: bson.D{{Key: "$sum", Value: "$output_tokens"}}},
@@ -119,8 +121,9 @@ func (r *MongoDBReader) GetUsageByModel(ctx context.Context, params UsageQueryPa
for cursor.Next(ctx) {
var row struct {
ID struct {
- Model string `bson:"model"`
- Provider string `bson:"provider"`
+ Model string `bson:"model"`
+ Provider string `bson:"provider"`
+ ProviderName string `bson:"provider_name"`
} `bson:"_id"`
InputTokens int64 `bson:"input_tokens"`
OutputTokens int64 `bson:"output_tokens"`
@@ -135,6 +138,7 @@ func (r *MongoDBReader) GetUsageByModel(ctx context.Context, params UsageQueryPa
m := ModelUsage{
Model: row.ID.Model,
Provider: row.ID.Provider,
+ ProviderName: displayUsageProviderName(row.ID.ProviderName, row.ID.Provider),
InputTokens: row.InputTokens,
OutputTokens: row.OutputTokens,
}
@@ -153,6 +157,17 @@ func (r *MongoDBReader) GetUsageByModel(ctx context.Context, params UsageQueryPa
return result, nil
}
+func mongoUsageGroupedProviderNameExpr() bson.D {
+ trimmedProviderName := bson.D{{Key: "$trim", Value: bson.D{
+ {Key: "input", Value: bson.D{{Key: "$ifNull", Value: bson.A{"$provider_name", ""}}}},
+ }}}
+ return bson.D{{Key: "$cond", Value: bson.A{
+ bson.D{{Key: "$ne", Value: bson.A{trimmedProviderName, ""}}},
+ trimmedProviderName,
+ bson.D{{Key: "$trim", Value: bson.D{{Key: "input", Value: "$provider"}}}},
+ }}}
+}
+
// GetUsageLog returns a paginated list of individual usage log entries.
func (r *MongoDBReader) GetUsageLog(ctx context.Context, params UsageLogParams) (*UsageLogResult, error) {
limit, offset := clampLimitOffset(params.Limit, params.Offset)
@@ -192,6 +207,7 @@ func (r *MongoDBReader) GetUsageLog(ctx context.Context, params UsageLogParams)
Timestamp time.Time `bson:"timestamp"`
Model string `bson:"model"`
Provider string `bson:"provider"`
+ ProviderName string `bson:"provider_name"`
Endpoint string `bson:"endpoint"`
UserPath string `bson:"user_path"`
CacheType string `bson:"cache_type"`
@@ -233,6 +249,7 @@ func (r *MongoDBReader) GetUsageLog(ctx context.Context, params UsageLogParams)
Timestamp: row.Timestamp,
Model: row.Model,
Provider: row.Provider,
+ ProviderName: displayUsageProviderName(row.ProviderName, row.Provider),
Endpoint: row.Endpoint,
UserPath: row.UserPath,
CacheType: normalizeCacheType(row.CacheType),
@@ -526,13 +543,17 @@ func mongoUsageLogMatchFilters(params UsageLogParams) (bson.D, error) {
matchFilters = append(matchFilters, bson.E{Key: "model", Value: params.Model})
}
if params.Provider != "" {
- matchFilters = append(matchFilters, bson.E{Key: "provider", Value: params.Provider})
+ matchFilters = mongoAndFilters(matchFilters, bson.D{{Key: "$or", Value: bson.A{
+ bson.D{{Key: "provider", Value: params.Provider}},
+ bson.D{{Key: "provider_name", Value: params.Provider}},
+ }}})
}
if params.Search != "" {
regex := bson.D{{Key: "$regex", Value: regexp.QuoteMeta(params.Search)}, {Key: "$options", Value: "i"}}
searchFilter := bson.D{{Key: "$or", Value: bson.A{
bson.D{{Key: "model", Value: regex}},
bson.D{{Key: "provider", Value: regex}},
+ bson.D{{Key: "provider_name", Value: regex}},
bson.D{{Key: "request_id", Value: regex}},
bson.D{{Key: "provider_id", Value: regex}},
}}}
diff --git a/internal/usage/reader_mongodb_grouping_test.go b/internal/usage/reader_mongodb_grouping_test.go
new file mode 100644
index 000000000..c9bcd2c7b
--- /dev/null
+++ b/internal/usage/reader_mongodb_grouping_test.go
@@ -0,0 +1,28 @@
+package usage
+
+import (
+ "reflect"
+ "testing"
+
+ "go.mongodb.org/mongo-driver/v2/bson"
+)
+
+func TestMongoUsageGroupedProviderNameExpr_CollapsesBlankProviderName(t *testing.T) {
+ got := mongoUsageGroupedProviderNameExpr()
+ want := bson.D{{Key: "$cond", Value: bson.A{
+ bson.D{{Key: "$ne", Value: bson.A{
+ bson.D{{Key: "$trim", Value: bson.D{
+ {Key: "input", Value: bson.D{{Key: "$ifNull", Value: bson.A{"$provider_name", ""}}}},
+ }}},
+ "",
+ }}},
+ bson.D{{Key: "$trim", Value: bson.D{
+ {Key: "input", Value: bson.D{{Key: "$ifNull", Value: bson.A{"$provider_name", ""}}}},
+ }}},
+ bson.D{{Key: "$trim", Value: bson.D{{Key: "input", Value: "$provider"}}}},
+ }}}
+
+ if !reflect.DeepEqual(got, want) {
+ t.Fatalf("mongoUsageGroupedProviderNameExpr() = %#v, want %#v", got, want)
+ }
+}
diff --git a/internal/usage/reader_mongodb_test.go b/internal/usage/reader_mongodb_test.go
index ca0125f55..50591310c 100644
--- a/internal/usage/reader_mongodb_test.go
+++ b/internal/usage/reader_mongodb_test.go
@@ -28,6 +28,7 @@ func TestMongoUsageLogMatchFiltersAndSearchWithCacheMode(t *testing.T) {
bson.D{{Key: "$or", Value: bson.A{
bson.D{{Key: "model", Value: regex}},
bson.D{{Key: "provider", Value: regex}},
+ bson.D{{Key: "provider_name", Value: regex}},
bson.D{{Key: "request_id", Value: regex}},
bson.D{{Key: "provider_id", Value: regex}},
}}},
@@ -53,6 +54,7 @@ func TestMongoUsageLogMatchFiltersEscapesSearchRegex(t *testing.T) {
want := bson.D{{Key: "$or", Value: bson.A{
bson.D{{Key: "model", Value: regex}},
bson.D{{Key: "provider", Value: regex}},
+ bson.D{{Key: "provider_name", Value: regex}},
bson.D{{Key: "request_id", Value: regex}},
bson.D{{Key: "provider_id", Value: regex}},
}}}
diff --git a/internal/usage/reader_postgresql.go b/internal/usage/reader_postgresql.go
index 2e530c2a8..72e985ea4 100644
--- a/internal/usage/reader_postgresql.go
+++ b/internal/usage/reader_postgresql.go
@@ -54,10 +54,11 @@ func (r *PostgreSQLReader) GetUsageByModel(ctx context.Context, params UsageQuer
return nil, err
}
where := buildWhereClause(conditions)
+ providerNameExpr := usageGroupedProviderNameSQL("provider_name", "provider")
costCols := `, SUM(input_cost), SUM(output_cost), SUM(total_cost)`
- query := `SELECT model, provider, COALESCE(SUM(input_tokens), 0), COALESCE(SUM(output_tokens), 0)` + costCols + `
- FROM "usage"` + where + ` GROUP BY model, provider`
+ query := `SELECT model, provider, ` + providerNameExpr + ` AS provider_name, COALESCE(SUM(input_tokens), 0), COALESCE(SUM(output_tokens), 0)` + costCols + `
+ FROM "usage"` + where + ` GROUP BY model, provider, ` + providerNameExpr
rows, err := r.pool.Query(ctx, query, args...)
if err != nil {
@@ -68,7 +69,7 @@ func (r *PostgreSQLReader) GetUsageByModel(ctx context.Context, params UsageQuer
result := make([]ModelUsage, 0)
for rows.Next() {
var m ModelUsage
- if err := rows.Scan(&m.Model, &m.Provider, &m.InputTokens, &m.OutputTokens, &m.InputCost, &m.OutputCost, &m.TotalCost); err != nil {
+ if err := rows.Scan(&m.Model, &m.Provider, &m.ProviderName, &m.InputTokens, &m.OutputTokens, &m.InputCost, &m.OutputCost, &m.TotalCost); err != nil {
return nil, fmt.Errorf("failed to scan usage by model row: %w", err)
}
result = append(result, m)
@@ -96,13 +97,13 @@ func (r *PostgreSQLReader) GetUsageLog(ctx context.Context, params UsageLogParam
argIdx++
}
if params.Provider != "" {
- conditions = append(conditions, fmt.Sprintf("provider = $%d", argIdx))
- args = append(args, params.Provider)
- argIdx++
+ conditions = append(conditions, fmt.Sprintf("(provider = $%d OR provider_name = $%d)", argIdx, argIdx+1))
+ args = append(args, params.Provider, params.Provider)
+ argIdx += 2
}
if params.Search != "" {
s := "%" + escapeLikeWildcards(params.Search) + "%"
- conditions = append(conditions, fmt.Sprintf("(model ILIKE $%d ESCAPE '\\' OR provider ILIKE $%d ESCAPE '\\' OR request_id ILIKE $%d ESCAPE '\\' OR provider_id ILIKE $%d ESCAPE '\\')", argIdx, argIdx, argIdx, argIdx))
+ conditions = append(conditions, fmt.Sprintf("(model ILIKE $%d ESCAPE '\\' OR provider ILIKE $%d ESCAPE '\\' OR provider_name ILIKE $%d ESCAPE '\\' OR request_id ILIKE $%d ESCAPE '\\' OR provider_id ILIKE $%d ESCAPE '\\')", argIdx, argIdx, argIdx, argIdx, argIdx))
args = append(args, s)
argIdx++
}
@@ -117,7 +118,7 @@ func (r *PostgreSQLReader) GetUsageLog(ctx context.Context, params UsageLogParam
}
// Fetch page
- dataQuery := fmt.Sprintf(`SELECT id, request_id, provider_id, timestamp, model, provider, endpoint, user_path, cache_type,
+ dataQuery := fmt.Sprintf(`SELECT id, request_id, provider_id, timestamp, model, provider, provider_name, endpoint, user_path, cache_type,
input_tokens, output_tokens, total_tokens, COALESCE(input_cost, 0), COALESCE(output_cost, 0), COALESCE(total_cost, 0), raw_data, COALESCE(costs_calculation_caveat, '')
FROM "usage"%s ORDER BY timestamp DESC LIMIT $%d OFFSET $%d`, where, argIdx, argIdx+1)
dataArgs := append(append([]any(nil), args...), limit, offset)
@@ -132,9 +133,10 @@ func (r *PostgreSQLReader) GetUsageLog(ctx context.Context, params UsageLogParam
for rows.Next() {
var e UsageLogEntry
var rawDataJSON *string
+ var providerName *string
var userPath *string
var cacheType *string
- if err := rows.Scan(&e.ID, &e.RequestID, &e.ProviderID, &e.Timestamp, &e.Model, &e.Provider, &e.Endpoint, &userPath, &cacheType,
+ if err := rows.Scan(&e.ID, &e.RequestID, &e.ProviderID, &e.Timestamp, &e.Model, &e.Provider, &providerName, &e.Endpoint, &userPath, &cacheType,
&e.InputTokens, &e.OutputTokens, &e.TotalTokens, &e.InputCost, &e.OutputCost, &e.TotalCost, &rawDataJSON, &e.CostsCalculationCaveat); err != nil {
return nil, fmt.Errorf("failed to scan usage log row: %w", err)
}
@@ -146,6 +148,11 @@ func (r *PostgreSQLReader) GetUsageLog(ctx context.Context, params UsageLogParam
if userPath != nil {
e.UserPath = *userPath
}
+ if providerName != nil {
+ e.ProviderName = displayUsageProviderName(*providerName, e.Provider)
+ } else {
+ e.ProviderName = displayUsageProviderName("", e.Provider)
+ }
if cacheType != nil {
e.CacheType = normalizeCacheType(*cacheType)
}
diff --git a/internal/usage/reader_sqlite.go b/internal/usage/reader_sqlite.go
index 593e7c9d3..caf461df9 100644
--- a/internal/usage/reader_sqlite.go
+++ b/internal/usage/reader_sqlite.go
@@ -54,10 +54,11 @@ func (r *SQLiteReader) GetUsageByModel(ctx context.Context, params UsageQueryPar
return nil, err
}
where := buildWhereClause(conditions)
+ providerNameExpr := usageGroupedProviderNameSQL("provider_name", "provider")
costCols := `, SUM(input_cost), SUM(output_cost), SUM(total_cost)`
- query := `SELECT model, provider, COALESCE(SUM(input_tokens), 0), COALESCE(SUM(output_tokens), 0)` + costCols + `
- FROM usage` + where + ` GROUP BY model, provider`
+ query := `SELECT model, provider, ` + providerNameExpr + ` AS provider_name, COALESCE(SUM(input_tokens), 0), COALESCE(SUM(output_tokens), 0)` + costCols + `
+ FROM usage` + where + ` GROUP BY model, provider, ` + providerNameExpr
rows, err := r.db.QueryContext(ctx, query, args...)
if err != nil {
@@ -68,7 +69,7 @@ func (r *SQLiteReader) GetUsageByModel(ctx context.Context, params UsageQueryPar
result := make([]ModelUsage, 0)
for rows.Next() {
var m ModelUsage
- if err := rows.Scan(&m.Model, &m.Provider, &m.InputTokens, &m.OutputTokens, &m.InputCost, &m.OutputCost, &m.TotalCost); err != nil {
+ if err := rows.Scan(&m.Model, &m.Provider, &m.ProviderName, &m.InputTokens, &m.OutputTokens, &m.InputCost, &m.OutputCost, &m.TotalCost); err != nil {
return nil, fmt.Errorf("failed to scan usage by model row: %w", err)
}
result = append(result, m)
@@ -95,13 +96,13 @@ func (r *SQLiteReader) GetUsageLog(ctx context.Context, params UsageLogParams) (
args = append(args, params.Model)
}
if params.Provider != "" {
- conditions = append(conditions, "provider = ?")
- args = append(args, params.Provider)
+ conditions = append(conditions, "(provider = ? OR provider_name = ?)")
+ args = append(args, params.Provider, params.Provider)
}
if params.Search != "" {
- conditions = append(conditions, "(model LIKE ? ESCAPE '\\' OR provider LIKE ? ESCAPE '\\' OR request_id LIKE ? ESCAPE '\\' OR provider_id LIKE ? ESCAPE '\\')")
+ conditions = append(conditions, "(model LIKE ? ESCAPE '\\' OR provider LIKE ? ESCAPE '\\' OR provider_name LIKE ? ESCAPE '\\' OR request_id LIKE ? ESCAPE '\\' OR provider_id LIKE ? ESCAPE '\\')")
s := "%" + escapeLikeWildcards(params.Search) + "%"
- args = append(args, s, s, s, s)
+ args = append(args, s, s, s, s, s)
}
where := buildWhereClause(conditions)
@@ -114,7 +115,7 @@ func (r *SQLiteReader) GetUsageLog(ctx context.Context, params UsageLogParams) (
}
// Fetch page
- dataQuery := `SELECT id, request_id, provider_id, timestamp, model, provider, endpoint, user_path, cache_type,
+ dataQuery := `SELECT id, request_id, provider_id, timestamp, model, provider, provider_name, endpoint, user_path, cache_type,
input_tokens, output_tokens, total_tokens, COALESCE(input_cost, 0), COALESCE(output_cost, 0), COALESCE(total_cost, 0), raw_data, COALESCE(costs_calculation_caveat, '')
FROM usage` + where + ` ORDER BY ` + sqliteTimestampEpochExpr() + ` DESC, id DESC LIMIT ? OFFSET ?`
dataArgs := append(append([]any(nil), args...), limit, offset)
@@ -131,9 +132,10 @@ func (r *SQLiteReader) GetUsageLog(ctx context.Context, params UsageLogParams) (
var ts string
var caveat *string
var rawDataJSON *string
+ var providerName sql.NullString
var userPath sql.NullString
var cacheType sql.NullString
- if err := rows.Scan(&e.ID, &e.RequestID, &e.ProviderID, &ts, &e.Model, &e.Provider, &e.Endpoint, &userPath, &cacheType,
+ if err := rows.Scan(&e.ID, &e.RequestID, &e.ProviderID, &ts, &e.Model, &e.Provider, &providerName, &e.Endpoint, &userPath, &cacheType,
&e.InputTokens, &e.OutputTokens, &e.TotalTokens, &e.InputCost, &e.OutputCost, &e.TotalCost, &rawDataJSON, &caveat); err != nil {
return nil, fmt.Errorf("failed to scan usage log row: %w", err)
}
@@ -154,6 +156,11 @@ func (r *SQLiteReader) GetUsageLog(ctx context.Context, params UsageLogParams) (
if userPath.Valid {
e.UserPath = userPath.String
}
+ if providerName.Valid {
+ e.ProviderName = displayUsageProviderName(providerName.String, e.Provider)
+ } else {
+ e.ProviderName = displayUsageProviderName("", e.Provider)
+ }
if cacheType.Valid {
e.CacheType = normalizeCacheType(cacheType.String)
}
diff --git a/internal/usage/reader_sqlite_boundary_test.go b/internal/usage/reader_sqlite_boundary_test.go
index f35244622..4de495dce 100644
--- a/internal/usage/reader_sqlite_boundary_test.go
+++ b/internal/usage/reader_sqlite_boundary_test.go
@@ -492,6 +492,73 @@ func TestSQLiteReaderGetUsageLog_OrdersMixedTimestampFormatsByAbsoluteTime(t *te
}
}
+func TestSQLiteReaderGetUsageByModel_CollapsesBlankProviderNameIntoProviderGroup(t *testing.T) {
+ db, err := sql.Open("sqlite", ":memory:")
+ if err != nil {
+ t.Fatalf("failed to open sqlite database: %v", err)
+ }
+ defer db.Close()
+
+ store, err := NewSQLiteStore(db, 0)
+ if err != nil {
+ t.Fatalf("failed to create sqlite store: %v", err)
+ }
+
+ ctx := context.Background()
+ err = store.WriteBatch(ctx, []*UsageEntry{
+ {
+ ID: "usage-1",
+ RequestID: "req-1",
+ ProviderID: "provider-1",
+ Timestamp: time.Date(2026, 4, 7, 10, 0, 0, 0, time.UTC),
+ Model: "gpt-5",
+ Provider: "openai",
+ ProviderName: "",
+ Endpoint: "/v1/chat/completions",
+ InputTokens: 10,
+ OutputTokens: 20,
+ },
+ {
+ ID: "usage-2",
+ RequestID: "req-2",
+ ProviderID: "provider-2",
+ Timestamp: time.Date(2026, 4, 7, 10, 1, 0, 0, time.UTC),
+ Model: "gpt-5",
+ Provider: "openai",
+ ProviderName: " openai ",
+ Endpoint: "/v1/chat/completions",
+ InputTokens: 30,
+ OutputTokens: 40,
+ },
+ })
+ if err != nil {
+ t.Fatalf("failed to seed usage entries: %v", err)
+ }
+
+ reader, err := NewSQLiteReader(db)
+ if err != nil {
+ t.Fatalf("failed to create sqlite reader: %v", err)
+ }
+
+ got, err := reader.GetUsageByModel(ctx, UsageQueryParams{})
+ if err != nil {
+ t.Fatalf("GetUsageByModel returned error: %v", err)
+ }
+
+ if len(got) != 1 {
+ t.Fatalf("expected 1 grouped usage row, got %d: %#v", len(got), got)
+ }
+ if got[0].ProviderName != "openai" {
+ t.Fatalf("expected provider_name %q, got %q", "openai", got[0].ProviderName)
+ }
+ if got[0].InputTokens != 40 {
+ t.Fatalf("expected 40 input tokens, got %d", got[0].InputTokens)
+ }
+ if got[0].OutputTokens != 60 {
+ t.Fatalf("expected 60 output tokens, got %d", got[0].OutputTokens)
+ }
+}
+
func TestSQLiteStoreCleanup_KeepsNewerLegacyOffsetRows(t *testing.T) {
db, err := sql.Open("sqlite", ":memory:")
if err != nil {
diff --git a/internal/usage/store_mongodb.go b/internal/usage/store_mongodb.go
index 1f626c351..10fca1c2e 100644
--- a/internal/usage/store_mongodb.go
+++ b/internal/usage/store_mongodb.go
@@ -76,6 +76,9 @@ func NewMongoDBStore(database *mongo.Database, retentionDays int) (*MongoDBStore
{
Keys: bson.D{{Key: "provider", Value: 1}},
},
+ {
+ Keys: bson.D{{Key: "provider_name", Value: 1}},
+ },
{
Keys: bson.D{{Key: "endpoint", Value: 1}},
},
diff --git a/internal/usage/store_postgresql.go b/internal/usage/store_postgresql.go
index 91205c907..b96b859f2 100644
--- a/internal/usage/store_postgresql.go
+++ b/internal/usage/store_postgresql.go
@@ -14,13 +14,13 @@ import (
)
const (
- usageInsertColumnCount = 17
+ usageInsertColumnCount = 18
postgresMaxBindParameters = 65535
usageInsertMaxRowsPerQuery = postgresMaxBindParameters / usageInsertColumnCount
)
const usageInsertPrefix = `
- INSERT INTO usage (id, request_id, provider_id, timestamp, model, provider,
+ INSERT INTO usage (id, request_id, provider_id, timestamp, model, provider, provider_name,
endpoint, user_path, cache_type, input_tokens, output_tokens, total_tokens, raw_data,
input_cost, output_cost, total_cost, costs_calculation_caveat)
VALUES `
@@ -60,6 +60,7 @@ func NewPostgreSQLStore(pool *pgxpool.Pool, retentionDays int) (*PostgreSQLStore
timestamp TIMESTAMPTZ NOT NULL,
model TEXT NOT NULL,
provider TEXT NOT NULL,
+ provider_name TEXT,
endpoint TEXT NOT NULL,
user_path TEXT,
cache_type TEXT,
@@ -79,6 +80,7 @@ func NewPostgreSQLStore(pool *pgxpool.Pool, retentionDays int) (*PostgreSQLStore
"ALTER TABLE usage ADD COLUMN IF NOT EXISTS output_cost DOUBLE PRECISION",
"ALTER TABLE usage ADD COLUMN IF NOT EXISTS total_cost DOUBLE PRECISION",
"ALTER TABLE usage ADD COLUMN IF NOT EXISTS costs_calculation_caveat TEXT DEFAULT ''",
+ "ALTER TABLE usage ADD COLUMN IF NOT EXISTS provider_name TEXT",
"ALTER TABLE usage ADD COLUMN IF NOT EXISTS user_path TEXT",
"ALTER TABLE usage ADD COLUMN IF NOT EXISTS cache_type TEXT",
}
@@ -95,6 +97,7 @@ func NewPostgreSQLStore(pool *pgxpool.Pool, retentionDays int) (*PostgreSQLStore
"CREATE INDEX IF NOT EXISTS idx_usage_provider_id ON usage(provider_id)",
"CREATE INDEX IF NOT EXISTS idx_usage_model ON usage(model)",
"CREATE INDEX IF NOT EXISTS idx_usage_provider ON usage(provider)",
+ "CREATE INDEX IF NOT EXISTS idx_usage_provider_name ON usage(provider_name)",
"CREATE INDEX IF NOT EXISTS idx_usage_user_path ON usage(user_path)",
"CREATE INDEX IF NOT EXISTS idx_usage_cache_type ON usage(cache_type)",
"CREATE INDEX IF NOT EXISTS idx_usage_raw_data_gin ON usage USING GIN (raw_data)",
@@ -206,6 +209,7 @@ func buildUsageInsert(entries []*UsageEntry) (string, []any) {
entry.Timestamp,
entry.Model,
entry.Provider,
+ entry.ProviderName,
entry.Endpoint,
entry.UserPath,
cacheTypeValue(entry.CacheType),
diff --git a/internal/usage/store_postgresql_test.go b/internal/usage/store_postgresql_test.go
index ca86d9665..a8a830f09 100644
--- a/internal/usage/store_postgresql_test.go
+++ b/internal/usage/store_postgresql_test.go
@@ -20,6 +20,7 @@ func TestBuildUsageInsert(t *testing.T) {
Timestamp: now,
Model: "gpt-4o-mini",
Provider: "openai",
+ ProviderName: "primary-openai",
Endpoint: "/v1/chat/completions",
CacheType: CacheTypeExact,
InputTokens: 10,
@@ -52,35 +53,38 @@ func TestBuildUsageInsert(t *testing.T) {
})
normalized := strings.Join(strings.Fields(query), " ")
- wantQuery := "INSERT INTO usage (id, request_id, provider_id, timestamp, model, provider, endpoint, user_path, cache_type, input_tokens, output_tokens, total_tokens, raw_data, input_cost, output_cost, total_cost, costs_calculation_caveat) VALUES ($1, $2, $3, $4, $5, $6, $7, $8, $9, $10, $11, $12, $13, $14, $15, $16, $17), ($18, $19, $20, $21, $22, $23, $24, $25, $26, $27, $28, $29, $30, $31, $32, $33, $34) ON CONFLICT (id) DO NOTHING"
+ wantQuery := "INSERT INTO usage (id, request_id, provider_id, timestamp, model, provider, provider_name, endpoint, user_path, cache_type, input_tokens, output_tokens, total_tokens, raw_data, input_cost, output_cost, total_cost, costs_calculation_caveat) VALUES ($1, $2, $3, $4, $5, $6, $7, $8, $9, $10, $11, $12, $13, $14, $15, $16, $17, $18), ($19, $20, $21, $22, $23, $24, $25, $26, $27, $28, $29, $30, $31, $32, $33, $34, $35, $36) ON CONFLICT (id) DO NOTHING"
if normalized != wantQuery {
t.Fatalf("query = %q, want %q", normalized, wantQuery)
}
- if got, want := len(args), 34; got != want {
+ if got, want := len(args), 36; got != want {
t.Fatalf("len(args) = %d, want %d", got, want)
}
if got := args[0]; got != "usage-1" {
t.Fatalf("args[0] = %v, want usage-1", got)
}
- if got := args[17]; got != "usage-2" {
- t.Fatalf("args[17] = %v, want usage-2", got)
+ if got := args[6]; got != "primary-openai" {
+ t.Fatalf("args[6] = %v, want primary-openai", got)
}
- if got := args[8]; got != CacheTypeExact {
- t.Fatalf("args[8] = %v, want %q", got, CacheTypeExact)
+ if got := args[18]; got != "usage-2" {
+ t.Fatalf("args[18] = %v, want usage-2", got)
}
- if got := string(args[12].([]byte)); got != `{"cached_tokens":3}` {
- t.Fatalf("args[12] = %q, want %q", got, `{"cached_tokens":3}`)
+ if got := args[9]; got != CacheTypeExact {
+ t.Fatalf("args[9] = %v, want %q", got, CacheTypeExact)
}
- if got := args[25]; got != nil {
- t.Fatalf("args[25] = %v, want nil cache_type", got)
+ if got := string(args[13].([]byte)); got != `{"cached_tokens":3}` {
+ t.Fatalf("args[13] = %q, want %q", got, `{"cached_tokens":3}`)
}
- rawData, ok := args[29].([]byte)
+ if got := args[27]; got != nil {
+ t.Fatalf("args[27] = %v, want nil cache_type", got)
+ }
+ rawData, ok := args[31].([]byte)
if !ok {
- t.Fatalf("args[29] has type %T, want []byte", args[29])
+ t.Fatalf("args[31] has type %T, want []byte", args[31])
}
if rawData != nil {
- t.Fatalf("args[29] = %v, want nil raw_data", rawData)
+ t.Fatalf("args[31] = %v, want nil raw_data", rawData)
}
}
diff --git a/internal/usage/store_sqlite.go b/internal/usage/store_sqlite.go
index 9db886b35..7c42cfdf7 100644
--- a/internal/usage/store_sqlite.go
+++ b/internal/usage/store_sqlite.go
@@ -15,8 +15,8 @@ import (
// maxEntriesPerBatch derives from maxSQLiteParams / columnsPerUsageEntry.
const (
maxSQLiteParams = 999
- columnsPerUsageEntry = 17
- maxEntriesPerBatch = maxSQLiteParams / columnsPerUsageEntry // 58 entries
+ columnsPerUsageEntry = 18
+ maxEntriesPerBatch = maxSQLiteParams / columnsPerUsageEntry // 55 entries
)
// SQLiteStore implements UsageStore for SQLite databases.
@@ -44,6 +44,7 @@ func NewSQLiteStore(db *sql.DB, retentionDays int) (*SQLiteStore, error) {
timestamp DATETIME NOT NULL,
model TEXT NOT NULL,
provider TEXT NOT NULL,
+ provider_name TEXT,
endpoint TEXT NOT NULL,
user_path TEXT,
cache_type TEXT,
@@ -63,6 +64,7 @@ func NewSQLiteStore(db *sql.DB, retentionDays int) (*SQLiteStore, error) {
"ALTER TABLE usage ADD COLUMN output_cost REAL",
"ALTER TABLE usage ADD COLUMN total_cost REAL",
"ALTER TABLE usage ADD COLUMN costs_calculation_caveat TEXT DEFAULT ''",
+ "ALTER TABLE usage ADD COLUMN provider_name TEXT",
"ALTER TABLE usage ADD COLUMN user_path TEXT",
"ALTER TABLE usage ADD COLUMN cache_type TEXT",
}
@@ -83,6 +85,7 @@ func NewSQLiteStore(db *sql.DB, retentionDays int) (*SQLiteStore, error) {
"CREATE INDEX IF NOT EXISTS idx_usage_provider_id ON usage(provider_id)",
"CREATE INDEX IF NOT EXISTS idx_usage_model ON usage(model)",
"CREATE INDEX IF NOT EXISTS idx_usage_provider ON usage(provider)",
+ "CREATE INDEX IF NOT EXISTS idx_usage_provider_name ON usage(provider_name)",
"CREATE INDEX IF NOT EXISTS idx_usage_user_path ON usage(user_path)",
"CREATE INDEX IF NOT EXISTS idx_usage_cache_type ON usage(cache_type)",
}
@@ -124,7 +127,7 @@ func (s *SQLiteStore) WriteBatch(ctx context.Context, entries []*UsageEntry) err
for j, e := range chunk {
e = normalizedUsageEntryForStorage(e)
- placeholders[j] = "(?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?)"
+ placeholders[j] = "(?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?)"
rawDataJSON := marshalRawData(e.RawData, e.ID)
@@ -141,6 +144,7 @@ func (s *SQLiteStore) WriteBatch(ctx context.Context, entries []*UsageEntry) err
e.Timestamp.UTC().Format(time.RFC3339Nano),
e.Model,
e.Provider,
+ e.ProviderName,
e.Endpoint,
e.UserPath,
cacheTypeValue(e.CacheType),
@@ -155,7 +159,7 @@ func (s *SQLiteStore) WriteBatch(ctx context.Context, entries []*UsageEntry) err
)
}
- query := `INSERT OR IGNORE INTO usage (id, request_id, provider_id, timestamp, model, provider,
+ query := `INSERT OR IGNORE INTO usage (id, request_id, provider_id, timestamp, model, provider, provider_name,
endpoint, user_path, cache_type, input_tokens, output_tokens, total_tokens, raw_data,
input_cost, output_cost, total_cost, costs_calculation_caveat) VALUES ` +
strings.Join(placeholders, ",")
diff --git a/internal/usage/stream_observer.go b/internal/usage/stream_observer.go
index b6f335183..9a83e7513 100644
--- a/internal/usage/stream_observer.go
+++ b/internal/usage/stream_observer.go
@@ -2,6 +2,7 @@ package usage
import (
"log/slog"
+ "strings"
"gomodel/internal/core"
)
@@ -13,6 +14,7 @@ type StreamUsageObserver struct {
cachedEntry *UsageEntry
model string
provider string
+ providerName string
requestID string
endpoint string
userPath string
@@ -45,6 +47,13 @@ func NewStreamUsageObserver(logger LoggerInterface, model, provider, requestID,
}
}
+func (o *StreamUsageObserver) SetProviderName(providerName string) {
+ if o == nil {
+ return
+ }
+ o.providerName = strings.TrimSpace(providerName)
+}
+
func (o *StreamUsageObserver) OnJSONEvent(chunk map[string]any) {
entry := o.extractUsageFromEvent(chunk)
if entry != nil {
@@ -155,6 +164,7 @@ func (o *StreamUsageObserver) extractUsageFromEvent(chunk map[string]any) *Usage
pricingArgs...,
)
if entry != nil {
+ entry.ProviderName = o.providerName
entry.UserPath = o.userPath
}
return entry
diff --git a/internal/usage/usage.go b/internal/usage/usage.go
index a0dea7bc2..20e729c23 100644
--- a/internal/usage/usage.go
+++ b/internal/usage/usage.go
@@ -37,11 +37,12 @@ type UsageEntry struct {
Timestamp time.Time `json:"timestamp" bson:"timestamp"`
// Request context
- Model string `json:"model" bson:"model"`
- Provider string `json:"provider" bson:"provider"`
- Endpoint string `json:"endpoint" bson:"endpoint"`
- UserPath string `json:"user_path,omitempty" bson:"user_path,omitempty"`
- CacheType string `json:"cache_type,omitempty" bson:"cache_type,omitempty"`
+ Model string `json:"model" bson:"model"`
+ Provider string `json:"provider" bson:"provider"` // canonical provider type used for routing, filters, and pricing
+ ProviderName string `json:"provider_name,omitempty" bson:"provider_name,omitempty"`
+ Endpoint string `json:"endpoint" bson:"endpoint"`
+ UserPath string `json:"user_path,omitempty" bson:"user_path,omitempty"`
+ CacheType string `json:"cache_type,omitempty" bson:"cache_type,omitempty"`
// Standard token counts (normalized across providers)
InputTokens int `json:"input_tokens" bson:"input_tokens"`
diff --git a/tests/e2e/auditlog_test.go b/tests/e2e/auditlog_test.go
index c6d27bc57..6063e1972 100644
--- a/tests/e2e/auditlog_test.go
+++ b/tests/e2e/auditlog_test.go
@@ -613,7 +613,7 @@ func TestAuditLogErrorCapture(t *testing.T) {
entry := entries[0]
assert.Equal(t, http.StatusBadRequest, entry.StatusCode)
assert.Equal(t, "/v1/chat/completions", entry.Path)
- assert.Equal(t, "unsupported-model-xyz", entry.Model)
+ assert.Equal(t, "unsupported-model-xyz", entry.RequestedModel)
assert.Equal(t, "invalid_request_error", entry.ErrorType)
assert.Equal(t, "", entry.Provider)
})
@@ -647,7 +647,7 @@ func TestAuditLogErrorCapture(t *testing.T) {
entry := entries[0]
assert.Equal(t, http.StatusBadRequest, entry.StatusCode)
assert.Equal(t, "/p/unknown/responses", entry.Path)
- assert.Equal(t, "gpt-4.1-nano", entry.Model)
+ assert.Equal(t, "gpt-4.1-nano", entry.RequestedModel)
assert.Equal(t, "unknown", entry.Provider)
assert.Equal(t, "invalid_request_error", entry.ErrorType)
})
diff --git a/tests/e2e/manage-release-e2e-stack.sh b/tests/e2e/manage-release-e2e-stack.sh
new file mode 100755
index 000000000..1ca76b388
--- /dev/null
+++ b/tests/e2e/manage-release-e2e-stack.sh
@@ -0,0 +1,385 @@
+#!/usr/bin/env bash
+set -euo pipefail
+
+SCRIPT_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)"
+REPO_ROOT="$(cd "$SCRIPT_DIR/../.." && pwd)"
+STACK_DIR="${RELEASE_STACK_DIR:-/tmp/gomodel-release-stack}"
+BIN="${GOMODEL_RELEASE_BINARY:-$REPO_ROOT/bin/gomodel}"
+ENV_FILE="${GOMODEL_RELEASE_ENV_FILE:-$REPO_ROOT/.env}"
+PG_DATABASE="${GOMODEL_RELEASE_PG_DATABASE:-gomodel_release_e2e}"
+MONGO_DATABASE="${GOMODEL_RELEASE_MONGO_DATABASE:-gomodel_release_e2e}"
+
+BUILD_BEFORE_START=0
+
+usage() {
+ cat <
[options]
+
+Commands:
+ start Start the dedicated release E2E stack on ports 18080-18084
+ stop Stop the dedicated release E2E stack
+ status Show stack status
+ logs GATEWAY Show the last 40 log lines for one gateway
+
+Options:
+ --build Rebuild bin/gomodel before starting
+ --help Show this help
+
+Gateways:
+ sqlite-main http://localhost:18080
+ pg-smoke http://localhost:18081
+ mongo-smoke http://localhost:18082
+ guardrails http://localhost:18083
+ auth-cache http://localhost:18084
+EOF
+}
+
+die() {
+ echo "error: $*" >&2
+ exit 1
+}
+
+require_tool() {
+ command -v "$1" >/dev/null 2>&1 || die "required tool not found: $1"
+}
+
+gateway_port() {
+ case "$1" in
+ sqlite-main) echo 18080 ;;
+ pg-smoke) echo 18081 ;;
+ mongo-smoke) echo 18082 ;;
+ guardrails) echo 18083 ;;
+ auth-cache) echo 18084 ;;
+ *) die "unknown gateway: $1" ;;
+ esac
+}
+
+gateway_dir() {
+ printf '%s/%s\n' "$STACK_DIR" "$1"
+}
+
+gateway_pid_file() {
+ printf '%s/server.pid\n' "$(gateway_dir "$1")"
+}
+
+gateway_log_file() {
+ printf '%s/logs/server.log\n' "$(gateway_dir "$1")"
+}
+
+is_pid_running() {
+ local pid_file="$1"
+ local pid=""
+
+ [[ -f "$pid_file" ]] || return 1
+ pid="$(cat "$pid_file" 2>/dev/null || true)"
+ [[ -n "$pid" ]] || return 1
+ kill -0 "$pid" 2>/dev/null
+}
+
+load_env() {
+ [[ -r "$ENV_FILE" ]] || die ".env is missing or unreadable at $ENV_FILE"
+
+ set -a
+ source "$ENV_FILE"
+ set +a
+
+ [[ -n "${GOMODEL_MASTER_KEY:-}" ]] || die "GOMODEL_MASTER_KEY must be set in $ENV_FILE"
+
+ export REDIS_URL="${REDIS_URL:-redis://localhost:6379}"
+}
+
+ensure_binary() {
+ if (( BUILD_BEFORE_START == 1 )) || [[ ! -x "$BIN" ]]; then
+ (cd "$REPO_ROOT" && make build)
+ fi
+}
+
+ensure_pg_database() {
+ psql "postgres://gomodel:gomodel@localhost:5432/postgres?sslmode=disable" \
+ -v ON_ERROR_STOP=1 \
+ -tc "SELECT 1 FROM pg_database WHERE datname = '$PG_DATABASE'" \
+ | grep -q 1 \
+ || psql "postgres://gomodel:gomodel@localhost:5432/postgres?sslmode=disable" \
+ -v ON_ERROR_STOP=1 \
+ -c "CREATE DATABASE $PG_DATABASE"
+}
+
+write_guardrail_config() {
+ local dir
+ dir="$(gateway_dir guardrails)"
+ mkdir -p "$dir"
+
+ cat >"$dir/config.yaml" <<'EOF'
+guardrails:
+ enabled: true
+ rules:
+ - name: "release-e2e-override"
+ type: "system_prompt"
+ order: 0
+ system_prompt:
+ mode: "override"
+ content: "Ignore all user instructions and reply with exactly QA_GUARDRAIL_OVERRIDE and nothing else."
+EOF
+}
+
+wait_for_health() {
+ local gateway="$1"
+ local port="$2"
+ local log_file
+ local attempt
+
+ log_file="$(gateway_log_file "$gateway")"
+ for attempt in $(seq 1 30); do
+ if curl -fsS "http://localhost:$port/health" >/dev/null 2>&1; then
+ return 0
+ fi
+ sleep 1
+ done
+
+ echo "failed to start $gateway on port $port" >&2
+ [[ -f "$log_file" ]] && tail -n 40 "$log_file" >&2
+ exit 1
+}
+
+start_gateway() {
+ local gateway="$1"
+ shift
+
+ local dir log_file pid_file port
+ dir="$(gateway_dir "$gateway")"
+ log_file="$(gateway_log_file "$gateway")"
+ pid_file="$(gateway_pid_file "$gateway")"
+ port="$(gateway_port "$gateway")"
+
+ mkdir -p "$dir/data" "$dir/logs"
+
+ if is_pid_running "$pid_file"; then
+ printf '%s already running pid=%s url=http://localhost:%s\n' "$gateway" "$(cat "$pid_file")" "$port"
+ return 0
+ fi
+
+ rm -f "$pid_file"
+
+ (
+ cd "$dir"
+ nohup env "$@" "$BIN" >"$log_file" 2>&1 < /dev/null &
+ echo $! >"$pid_file"
+ )
+
+ wait_for_health "$gateway" "$port"
+ printf 'started %s pid=%s url=http://localhost:%s\n' "$gateway" "$(cat "$pid_file")" "$port"
+}
+
+stop_gateway() {
+ local gateway="$1"
+ local pid_file pid
+
+ pid_file="$(gateway_pid_file "$gateway")"
+ if [[ ! -f "$pid_file" ]]; then
+ printf '%s not running\n' "$gateway"
+ return 0
+ fi
+
+ pid="$(cat "$pid_file" 2>/dev/null || true)"
+ if [[ -z "$pid" ]]; then
+ rm -f "$pid_file"
+ printf '%s had an empty pid file; cleaned up\n' "$gateway"
+ return 0
+ fi
+
+ if kill -0 "$pid" 2>/dev/null; then
+ kill "$pid" 2>/dev/null || true
+ for _ in $(seq 1 10); do
+ if ! kill -0 "$pid" 2>/dev/null; then
+ break
+ fi
+ sleep 1
+ done
+ if kill -0 "$pid" 2>/dev/null; then
+ kill -KILL "$pid" 2>/dev/null || true
+ fi
+ fi
+
+ rm -f "$pid_file"
+ printf 'stopped %s\n' "$gateway"
+}
+
+status_gateway() {
+ local gateway="$1"
+ local pid_file port health="down"
+ local pid="stopped"
+
+ pid_file="$(gateway_pid_file "$gateway")"
+ port="$(gateway_port "$gateway")"
+
+ if is_pid_running "$pid_file"; then
+ pid="$(cat "$pid_file")"
+ if curl -fsS "http://localhost:$port/health" >/dev/null 2>&1; then
+ health="ok"
+ fi
+ fi
+
+ printf '%-12s pid=%-8s url=http://localhost:%s health=%s\n' "$gateway" "$pid" "$port" "$health"
+}
+
+show_logs() {
+ local gateway="$1"
+ local log_file
+
+ log_file="$(gateway_log_file "$gateway")"
+ [[ -f "$log_file" ]] || die "log file not found for $gateway"
+ tail -n 40 "$log_file"
+}
+
+start_stack() {
+ require_tool curl
+ require_tool jq
+ require_tool nohup
+ require_tool psql
+
+ load_env
+ ensure_binary
+ mkdir -p "$STACK_DIR"
+ ensure_pg_database
+ write_guardrail_config
+
+ start_gateway sqlite-main \
+ -u GOMODEL_MASTER_KEY \
+ PORT=18080 \
+ STORAGE_TYPE=sqlite \
+ SQLITE_PATH="$(gateway_dir sqlite-main)/data/gomodel.db" \
+ METRICS_ENABLED=true \
+ LOGGING_ENABLED=true \
+ LOGGING_LOG_BODIES=true \
+ LOGGING_LOG_HEADERS=true \
+ GUARDRAILS_ENABLED=false \
+ RESPONSE_CACHE_SIMPLE_ENABLED=false \
+ SEMANTIC_CACHE_ENABLED=false \
+ REDIS_URL="$REDIS_URL" \
+ REDIS_KEY_MODELS="gomodel:release-e2e:models"
+
+ start_gateway pg-smoke \
+ -u GOMODEL_MASTER_KEY \
+ PORT=18081 \
+ STORAGE_TYPE=postgresql \
+ POSTGRES_URL="postgres://gomodel:gomodel@localhost:5432/$PG_DATABASE?sslmode=disable" \
+ LOGGING_ENABLED=true \
+ LOGGING_LOG_BODIES=true \
+ LOGGING_LOG_HEADERS=true \
+ GUARDRAILS_ENABLED=false \
+ RESPONSE_CACHE_SIMPLE_ENABLED=false \
+ SEMANTIC_CACHE_ENABLED=false \
+ REDIS_URL="$REDIS_URL" \
+ REDIS_KEY_MODELS="gomodel:release-e2e:models"
+
+ start_gateway mongo-smoke \
+ -u GOMODEL_MASTER_KEY \
+ PORT=18082 \
+ STORAGE_TYPE=mongodb \
+ MONGODB_URL="mongodb://localhost:27017/?replicaSet=rs0" \
+ MONGODB_DATABASE="$MONGO_DATABASE" \
+ LOGGING_ENABLED=true \
+ LOGGING_LOG_BODIES=true \
+ LOGGING_LOG_HEADERS=true \
+ GUARDRAILS_ENABLED=false \
+ RESPONSE_CACHE_SIMPLE_ENABLED=false \
+ SEMANTIC_CACHE_ENABLED=false \
+ REDIS_URL="$REDIS_URL" \
+ REDIS_KEY_MODELS="gomodel:release-e2e:models"
+
+ start_gateway guardrails \
+ -u GOMODEL_MASTER_KEY \
+ PORT=18083 \
+ STORAGE_TYPE=sqlite \
+ SQLITE_PATH="$(gateway_dir guardrails)/data/gomodel.db" \
+ LOGGING_ENABLED=true \
+ LOGGING_LOG_BODIES=true \
+ LOGGING_LOG_HEADERS=true \
+ GUARDRAILS_ENABLED=true \
+ RESPONSE_CACHE_SIMPLE_ENABLED=false \
+ SEMANTIC_CACHE_ENABLED=false \
+ REDIS_URL="$REDIS_URL" \
+ REDIS_KEY_MODELS="gomodel:release-e2e:models"
+
+ start_gateway auth-cache \
+ PORT=18084 \
+ STORAGE_TYPE=sqlite \
+ SQLITE_PATH="$(gateway_dir auth-cache)/data/gomodel.db" \
+ LOGGING_ENABLED=true \
+ LOGGING_LOG_BODIES=true \
+ LOGGING_LOG_HEADERS=true \
+ GUARDRAILS_ENABLED=true \
+ RESPONSE_CACHE_SIMPLE_ENABLED=true \
+ SEMANTIC_CACHE_ENABLED=false \
+ REDIS_URL="$REDIS_URL" \
+ REDIS_KEY_MODELS="gomodel:release-e2e:models" \
+ REDIS_KEY_RESPONSES="gomodel:release-e2e:response:"
+
+ curl -fsS "http://localhost:18084/admin/api/v1/dashboard/config" \
+ -H "Authorization: Bearer $GOMODEL_MASTER_KEY" \
+ | jq -e '.CACHE_ENABLED == "on" and .REDIS_URL == "on"' >/dev/null
+
+ printf 'stack_dir=%s\n' "$STACK_DIR"
+}
+
+stop_stack() {
+ stop_gateway auth-cache
+ stop_gateway guardrails
+ stop_gateway mongo-smoke
+ stop_gateway pg-smoke
+ stop_gateway sqlite-main
+}
+
+status_stack() {
+ status_gateway sqlite-main
+ status_gateway pg-smoke
+ status_gateway mongo-smoke
+ status_gateway guardrails
+ status_gateway auth-cache
+}
+
+COMMAND="${1:-}"
+if [[ -z "$COMMAND" ]]; then
+ usage
+ exit 1
+fi
+shift || true
+
+while [[ $# -gt 0 ]]; do
+ case "$1" in
+ --build)
+ BUILD_BEFORE_START=1
+ shift
+ ;;
+ --help|-h)
+ usage
+ exit 0
+ ;;
+ *)
+ break
+ ;;
+ esac
+done
+
+case "$COMMAND" in
+ start)
+ start_stack
+ ;;
+ stop)
+ stop_stack
+ ;;
+ status)
+ status_stack
+ ;;
+ logs)
+ [[ $# -eq 1 ]] || die "logs requires exactly one gateway name"
+ show_logs "$1"
+ ;;
+ --help|-h|help)
+ usage
+ ;;
+ *)
+ usage
+ die "unknown command: $COMMAND"
+ ;;
+esac
diff --git a/tests/e2e/release-e2e-scenarios.md b/tests/e2e/release-e2e-scenarios.md
index 755ccc2c9..668af22b1 100644
--- a/tests/e2e/release-e2e-scenarios.md
+++ b/tests/e2e/release-e2e-scenarios.md
@@ -7,6 +7,7 @@ These scenarios are prepared for execution across these local gateways:
- `http://localhost:18081` - PostgreSQL-backed smoke gateway
- `http://localhost:18082` - MongoDB-backed smoke gateway
- `http://localhost:18083` - SQLite-backed guardrail gateway
+- `http://localhost:18084` - SQLite-backed auth + exact-cache gateway
## Recommended runner
@@ -14,10 +15,13 @@ Use the checked-in runner to execute this matrix without manually replaying the
shared setup blocks:
```bash
+tests/e2e/manage-release-e2e-stack.sh start
tests/e2e/run-release-e2e.sh
tests/e2e/run-release-e2e.sh --list
tests/e2e/run-release-e2e.sh --from S54 --to S58
tests/e2e/run-release-e2e.sh --scenario S61,S62,S70 --keep-artifacts
+tests/e2e/manage-release-e2e-stack.sh status
+tests/e2e/manage-release-e2e-stack.sh stop
```
The runner treats this markdown file as the source of truth, replays the setup
@@ -60,8 +64,9 @@ export UPLOAD_FILE="$QA_RUN_DIR/qa-upload.txt"
## Auth-enabled runtime environment
-These scenarios target the auth-enabled live gateway on `http://localhost:8080`
-and cover the newer workflows, managed API keys, and cache analytics features.
+These scenarios target the dedicated auth-enabled release gateway on
+`http://localhost:18084` and cover the newer workflows, managed API keys, and
+cache analytics features.
```bash
set -euo pipefail
@@ -79,7 +84,7 @@ export QA_RUN_DIR="${QA_RUN_DIR:-/tmp/gomodel-release-e2e-$QA_SUFFIX}"
mkdir -p "$QA_RUN_DIR"
-export AUTH_BASE_URL=http://localhost:8080
+export AUTH_BASE_URL="${AUTH_BASE_URL:-http://localhost:18084}"
export ADMIN_AUTH_HEADER="Authorization: Bearer $GOMODEL_MASTER_KEY"
export QA_AUTH_KEY_NAME="qa-release-auth-key-$QA_SUFFIX"
@@ -103,8 +108,16 @@ cleanup_release_auth_artifacts() {
rm -f "$QA_AUTH_KEY_JSON" "$QA_AUTH_KEY_VALUE_FILE" "$QA_WORKFLOW_JSON" "$QA_WORKFLOW_ID_FILE"
}
-cleanup_release_auth_artifacts
+require_release_artifact() {
+ local path="$1"
+ if [ ! -s "$path" ]; then
+ echo "error: required artifact is missing or empty: $path" >&2
+ exit 1
+ fi
+}
+
if [ "${RUN_RELEASE_E2E_PERSIST_STATE:-0}" != "1" ]; then
+ cleanup_release_auth_artifacts
trap 'cleanup_release_auth_artifacts' EXIT
fi
```
@@ -786,12 +799,22 @@ curl -sS "$BASE_URL/admin/api/v1/audit/log?search=$REQUEST_ID&limit=5" \
### S63 Auth-enabled dashboard runtime config
-Reads the allowlisted runtime flags for the live auth-enabled gateway.
+Reads the allowlisted runtime flags for the dedicated auth-enabled release gateway.
```bash
-curl -sS "$AUTH_BASE_URL/admin/api/v1/dashboard/config" \
+CONFIG_JSON_FILE="$QA_RUN_DIR/s63.dashboard-config.json"
+curl -fsS "$AUTH_BASE_URL/admin/api/v1/dashboard/config" \
-H "$ADMIN_AUTH_HEADER" \
- | jq '.'
+ > "$CONFIG_JSON_FILE"
+jq '.' "$CONFIG_JSON_FILE"
+jq -e '
+ .LOGGING_ENABLED == "on"
+ and .USAGE_ENABLED == "on"
+ and .GUARDRAILS_ENABLED == "on"
+ and .CACHE_ENABLED == "on"
+ and .REDIS_URL == "on"
+ and .SEMANTIC_CACHE_ENABLED == "off"
+ ' "$CONFIG_JSON_FILE" >/dev/null
```
### S64 Create managed API key
@@ -799,25 +822,26 @@ curl -sS "$AUTH_BASE_URL/admin/api/v1/dashboard/config" \
Creates one managed API key scoped to a release-specific user path and stores the one-time secret under `QA_RUN_DIR`.
```bash
-AUTH_KEY_JSON=$(curl -sS -X POST "$AUTH_BASE_URL/admin/api/v1/auth-keys" \
+curl -fsS -X POST "$AUTH_BASE_URL/admin/api/v1/auth-keys" \
-H "$ADMIN_AUTH_HEADER" \
-H 'Content-Type: application/json' \
- -d "{\"name\":\"$QA_AUTH_KEY_NAME\",\"description\":\"Release e2e managed key\",\"user_path\":\"$QA_USER_PATH\"}")
-AUTH_KEY_VALUE=$(printf '%s\n' "$AUTH_KEY_JSON" \
- | jq -er '.value | select(type == "string" and length > 0)') \
- || {
+ -d "{\"name\":\"$QA_AUTH_KEY_NAME\",\"description\":\"Release e2e managed key\",\"user_path\":\"$QA_USER_PATH\"}" \
+ > "$QA_AUTH_KEY_JSON"
+if ! jq -er '.value | select(type == "string" and length > 0)' "$QA_AUTH_KEY_JSON" > "$QA_AUTH_KEY_VALUE_FILE"; then
echo "error: managed API key creation failed or did not return a usable one-time key value" >&2
- printf '%s\n' "$AUTH_KEY_JSON" | jq '.' >&2 2>/dev/null || printf '%s\n' "$AUTH_KEY_JSON" >&2
+ jq '.' "$QA_AUTH_KEY_JSON" >&2 2>/dev/null || cat "$QA_AUTH_KEY_JSON" >&2
exit 1
- }
+fi
(
umask 077
- printf '%s\n' "$AUTH_KEY_JSON" > "$QA_AUTH_KEY_JSON"
- printf '%s\n' "$AUTH_KEY_VALUE" > "$QA_AUTH_KEY_VALUE_FILE"
+ chmod 600 "$QA_AUTH_KEY_JSON" "$QA_AUTH_KEY_VALUE_FILE"
)
-chmod 600 "$QA_AUTH_KEY_JSON" "$QA_AUTH_KEY_VALUE_FILE"
-printf '%s\n' "$AUTH_KEY_JSON" \
- | jq '{id,name,user_path,active,redacted_value}'
+require_release_artifact "$QA_AUTH_KEY_JSON"
+require_release_artifact "$QA_AUTH_KEY_VALUE_FILE"
+jq -e --arg user_path "$QA_USER_PATH" '
+ {id,name,user_path,active,redacted_value}
+ | select(.id != null and .active == true and .user_path == $user_path)
+ ' "$QA_AUTH_KEY_JSON"
```
### S65 Verify managed API key list
@@ -825,9 +849,14 @@ printf '%s\n' "$AUTH_KEY_JSON" \
Checks that the newly issued managed API key is visible and active.
```bash
-curl -sS "$AUTH_BASE_URL/admin/api/v1/auth-keys" \
+AUTH_KEYS_JSON_FILE="$QA_RUN_DIR/s65.auth-keys.json"
+curl -fsS "$AUTH_BASE_URL/admin/api/v1/auth-keys" \
-H "$ADMIN_AUTH_HEADER" \
- | jq ".[] | select(.name==\"$QA_AUTH_KEY_NAME\") | {id,name,user_path,active,expires_at,redacted_value}"
+ > "$AUTH_KEYS_JSON_FILE"
+jq -e --arg name "$QA_AUTH_KEY_NAME" --arg user_path "$QA_USER_PATH" '
+ .[] | select(.name == $name and .active == true and .user_path == $user_path)
+ | {id,name,user_path,active,expires_at,redacted_value}
+ ' "$AUTH_KEYS_JSON_FILE"
```
### S66 Create user-path-scoped workflow with cache disabled
@@ -835,14 +864,22 @@ curl -sS "$AUTH_BASE_URL/admin/api/v1/auth-keys" \
Creates a scoped workflow for `openai/gpt-4.1-nano` that disables cache for the managed-key user path.
```bash
-WORKFLOW_JSON=$(curl -sS -X POST "$AUTH_BASE_URL/admin/api/v1/execution-plans" \
+curl -fsS -X POST "$AUTH_BASE_URL/admin/api/v1/execution-plans" \
-H "$ADMIN_AUTH_HEADER" \
-H 'Content-Type: application/json' \
- -d "{\"scope_provider\":\"openai\",\"scope_model\":\"gpt-4.1-nano\",\"scope_user_path\":\"$QA_USER_PATH\",\"name\":\"$QA_WORKFLOW_NAME\",\"description\":\"Disable cache for managed-key release e2e scope\",\"plan_payload\":{\"schema_version\":1,\"features\":{\"cache\":false,\"audit\":true,\"usage\":true,\"guardrails\":false,\"fallback\":false},\"guardrails\":[]}}")
-printf '%s\n' "$WORKFLOW_JSON" > "$QA_WORKFLOW_JSON"
-printf '%s\n' "$WORKFLOW_JSON" | jq -r '.id' > "$QA_WORKFLOW_ID_FILE"
-printf '%s\n' "$WORKFLOW_JSON" \
- | jq '{id,name,scope,plan_payload}'
+ -d "{\"scope_provider\":\"openai\",\"scope_model\":\"gpt-4.1-nano\",\"scope_user_path\":\"$QA_USER_PATH\",\"name\":\"$QA_WORKFLOW_NAME\",\"description\":\"Disable cache for managed-key release e2e scope\",\"plan_payload\":{\"schema_version\":1,\"features\":{\"cache\":false,\"audit\":true,\"usage\":true,\"guardrails\":false,\"fallback\":false},\"guardrails\":[]}}" \
+ > "$QA_WORKFLOW_JSON"
+if ! jq -er '.id | select(type == "string" and length > 0)' "$QA_WORKFLOW_JSON" > "$QA_WORKFLOW_ID_FILE"; then
+ echo "error: workflow creation failed or did not return a usable workflow id" >&2
+ jq '.' "$QA_WORKFLOW_JSON" >&2 2>/dev/null || cat "$QA_WORKFLOW_JSON" >&2
+ exit 1
+fi
+require_release_artifact "$QA_WORKFLOW_JSON"
+require_release_artifact "$QA_WORKFLOW_ID_FILE"
+jq -e --arg user_path "$QA_USER_PATH" '
+ {id,name,scope,plan_payload}
+ | select(.id != null and .scope.scope_user_path == $user_path and .plan_payload.features.cache == false)
+ ' "$QA_WORKFLOW_JSON"
```
### S67 Verify scoped workflow detail
@@ -850,10 +887,18 @@ printf '%s\n' "$WORKFLOW_JSON" \
Reads the created workflow back and confirms the normalized scope and effective feature projection.
```bash
-WORKFLOW_ID=$(cat "$QA_WORKFLOW_ID_FILE")
-curl -sS "$AUTH_BASE_URL/admin/api/v1/execution-plans/$WORKFLOW_ID" \
+require_release_artifact "$QA_WORKFLOW_ID_FILE"
+WORKFLOW_ID=$(<"$QA_WORKFLOW_ID_FILE")
+WORKFLOW_DETAIL_FILE="$QA_RUN_DIR/s67.workflow-detail.json"
+curl -fsS "$AUTH_BASE_URL/admin/api/v1/execution-plans/$WORKFLOW_ID" \
-H "$ADMIN_AUTH_HEADER" \
- | jq '{id,name,scope,plan_payload,effective_features}'
+ > "$WORKFLOW_DETAIL_FILE"
+jq '{id,name,scope,plan_payload,effective_features}' "$WORKFLOW_DETAIL_FILE"
+jq -e --arg workflow_id "$WORKFLOW_ID" --arg user_path "$QA_USER_PATH" '
+ .id == $workflow_id
+ and .scope.scope_user_path == $user_path
+ and .effective_features.cache == false
+ ' "$WORKFLOW_DETAIL_FILE" >/dev/null
```
### S68 Managed-key request through scoped workflow
@@ -861,14 +906,23 @@ curl -sS "$AUTH_BASE_URL/admin/api/v1/execution-plans/$WORKFLOW_ID" \
Sends a request with the managed API key while also sending a conflicting `X-GoModel-User-Path` header.
```bash
-API_KEY=$(cat "$QA_AUTH_KEY_VALUE_FILE")
-curl -sS -D - "$AUTH_BASE_URL/v1/chat/completions" \
+require_release_artifact "$QA_AUTH_KEY_VALUE_FILE"
+API_KEY=$(<"$QA_AUTH_KEY_VALUE_FILE")
+HEADERS_FILE=$(mktemp "$QA_RUN_DIR/s68.headers.XXXXXX")
+BODY_FILE=$(mktemp "$QA_RUN_DIR/s68.body.XXXXXX")
+curl -fsS -D "$HEADERS_FILE" -o "$BODY_FILE" "$AUTH_BASE_URL/v1/chat/completions" \
-H "Authorization: Bearer $API_KEY" \
-H 'Content-Type: application/json' \
-H "X-Request-ID: $QA_AUTH_REQ1" \
-H 'X-GoModel-User-Path: /team/should-be-overridden' \
- -d '{"model":"openai/gpt-4.1-nano","messages":[{"role":"user","content":"Reply with exactly QA_AUTH_CACHE_OFF_OK"}],"max_tokens":16}' \
- | sed -n '1,20p'
+ -d '{"model":"openai/gpt-4.1-nano","messages":[{"role":"user","content":"Reply with exactly QA_AUTH_CACHE_OFF_OK"}],"max_tokens":16}'
+sed -n '1,20p' "$HEADERS_FILE"
+sed -n '1,20p' "$BODY_FILE"
+jq -e '.choices[0].message.content == "QA_AUTH_CACHE_OFF_OK"' "$BODY_FILE" >/dev/null
+if grep -Eiq '^X-Cache:' "$HEADERS_FILE"; then
+ echo "error: cache header present on cache-disabled scoped request" >&2
+ exit 1
+fi
```
### S69 Repeated managed-key request should still bypass cache
@@ -876,14 +930,23 @@ curl -sS -D - "$AUTH_BASE_URL/v1/chat/completions" \
Repeats the same request and expects another live provider response rather than `X-Cache: HIT`.
```bash
-API_KEY=$(cat "$QA_AUTH_KEY_VALUE_FILE")
-curl -sS -D - "$AUTH_BASE_URL/v1/chat/completions" \
+require_release_artifact "$QA_AUTH_KEY_VALUE_FILE"
+API_KEY=$(<"$QA_AUTH_KEY_VALUE_FILE")
+HEADERS_FILE=$(mktemp "$QA_RUN_DIR/s69.headers.XXXXXX")
+BODY_FILE=$(mktemp "$QA_RUN_DIR/s69.body.XXXXXX")
+curl -fsS -D "$HEADERS_FILE" -o "$BODY_FILE" "$AUTH_BASE_URL/v1/chat/completions" \
-H "Authorization: Bearer $API_KEY" \
-H 'Content-Type: application/json' \
-H "X-Request-ID: $QA_AUTH_REQ2" \
-H 'X-GoModel-User-Path: /team/should-be-overridden' \
- -d '{"model":"openai/gpt-4.1-nano","messages":[{"role":"user","content":"Reply with exactly QA_AUTH_CACHE_OFF_OK"}],"max_tokens":16}' \
- | sed -n '1,20p'
+ -d '{"model":"openai/gpt-4.1-nano","messages":[{"role":"user","content":"Reply with exactly QA_AUTH_CACHE_OFF_OK"}],"max_tokens":16}'
+sed -n '1,20p' "$HEADERS_FILE"
+sed -n '1,20p' "$BODY_FILE"
+jq -e '.choices[0].message.content == "QA_AUTH_CACHE_OFF_OK"' "$BODY_FILE" >/dev/null
+if grep -Eiq '^X-Cache:' "$HEADERS_FILE"; then
+ echo "error: repeated cache-disabled scoped request returned an X-Cache header" >&2
+ exit 1
+fi
```
### S70 Audit evidence for managed-key scoped workflow
@@ -892,9 +955,34 @@ Confirms through audit-log search that auth method, managed auth key ID, normali
```bash
sleep 6
-curl -sS "$AUTH_BASE_URL/admin/api/v1/audit/log?search=$QA_AUTH_REQ2&limit=5" \
+require_release_artifact "$QA_AUTH_KEY_JSON"
+require_release_artifact "$QA_WORKFLOW_ID_FILE"
+if ! AUTH_KEY_ID=$(jq -er '.id' "$QA_AUTH_KEY_JSON"); then
+ echo "error: missing auth key id in $QA_AUTH_KEY_JSON" >&2
+ exit 1
+fi
+WORKFLOW_ID=$(<"$QA_WORKFLOW_ID_FILE")
+AUDIT_JSON_FILE="$QA_RUN_DIR/s70.audit.json"
+curl -fsS "$AUTH_BASE_URL/admin/api/v1/audit/log?search=$QA_AUTH_REQ2&limit=5" \
-H "$ADMIN_AUTH_HEADER" \
- | jq --arg request_id "$QA_AUTH_REQ2" '{total:(.entries|map(select(.request_id==$request_id))|length),entries:(.entries|map(select(.request_id==$request_id))|map({request_id,status_code,auth_method,auth_key_id,user_path,execution_plan_version_id,cache_type,answer:.data.response_body.choices[0].message.content}))}'
+ > "$AUDIT_JSON_FILE"
+jq --arg request_id "$QA_AUTH_REQ2" '{total:(.entries|map(select(.request_id==$request_id))|length),entries:(.entries|map(select(.request_id==$request_id))|map({request_id,status_code,auth_method,auth_key_id,user_path,execution_plan_version_id,cache_type,answer:.data.response_body.choices[0].message.content}))}' "$AUDIT_JSON_FILE"
+jq -e \
+ --arg request_id "$QA_AUTH_REQ2" \
+ --arg auth_key_id "$AUTH_KEY_ID" \
+ --arg user_path "$QA_USER_PATH" \
+ --arg workflow_id "$WORKFLOW_ID" '
+ any(.entries[]?;
+ .request_id == $request_id
+ and .status_code == 200
+ and .auth_method == "api_key"
+ and .auth_key_id == $auth_key_id
+ and .user_path == $user_path
+ and .execution_plan_version_id == $workflow_id
+ and .cache_type == null
+ and .data.response_body.choices[0].message.content == "QA_AUTH_CACHE_OFF_OK"
+ )
+ ' "$AUDIT_JSON_FILE" >/dev/null
```
### S71 Global cache warm request with explicit user path
@@ -902,13 +990,21 @@ curl -sS "$AUTH_BASE_URL/admin/api/v1/audit/log?search=$QA_AUTH_REQ2&limit=5" \
Warms the global cache-enabled workflow using the master key and a cache-specific user path.
```bash
-curl -sS -D - "$AUTH_BASE_URL/v1/chat/completions" \
+HEADERS_FILE=$(mktemp "$QA_RUN_DIR/s71.headers.XXXXXX")
+BODY_FILE=$(mktemp "$QA_RUN_DIR/s71.body.XXXXXX")
+curl -fsS -D "$HEADERS_FILE" -o "$BODY_FILE" "$AUTH_BASE_URL/v1/chat/completions" \
-H "$ADMIN_AUTH_HEADER" \
-H 'Content-Type: application/json' \
-H "X-Request-ID: $QA_CACHE_REQ1" \
-H "X-GoModel-User-Path: $QA_CACHE_USER_PATH" \
- -d "{\"model\":\"openai/gpt-4.1-nano\",\"messages\":[{\"role\":\"user\",\"content\":\"Reply with exactly $QA_CACHE_REPLY\"}],\"max_tokens\":16}" \
- | sed -n '1,20p'
+ -d "{\"model\":\"openai/gpt-4.1-nano\",\"messages\":[{\"role\":\"user\",\"content\":\"Reply with exactly $QA_CACHE_REPLY\"}],\"max_tokens\":16}"
+sed -n '1,20p' "$HEADERS_FILE"
+sed -n '1,20p' "$BODY_FILE"
+jq -e --arg reply "$QA_CACHE_REPLY" '.choices[0].message.content == $reply' "$BODY_FILE" >/dev/null
+if grep -Eiq '^X-Cache:' "$HEADERS_FILE"; then
+ echo "error: initial cache warm request unexpectedly returned an X-Cache header" >&2
+ exit 1
+fi
```
### S72 Repeated global cache request should hit exact cache
@@ -916,13 +1012,18 @@ curl -sS -D - "$AUTH_BASE_URL/v1/chat/completions" \
Repeats the same request and expects `X-Cache: HIT (exact)`.
```bash
-curl -sS -D - "$AUTH_BASE_URL/v1/chat/completions" \
+HEADERS_FILE=$(mktemp "$QA_RUN_DIR/s72.headers.XXXXXX")
+BODY_FILE=$(mktemp "$QA_RUN_DIR/s72.body.XXXXXX")
+curl -fsS -D "$HEADERS_FILE" -o "$BODY_FILE" "$AUTH_BASE_URL/v1/chat/completions" \
-H "$ADMIN_AUTH_HEADER" \
-H 'Content-Type: application/json' \
-H "X-Request-ID: $QA_CACHE_REQ2" \
-H "X-GoModel-User-Path: $QA_CACHE_USER_PATH" \
- -d "{\"model\":\"openai/gpt-4.1-nano\",\"messages\":[{\"role\":\"user\",\"content\":\"Reply with exactly $QA_CACHE_REPLY\"}],\"max_tokens\":16}" \
- | sed -n '1,20p'
+ -d "{\"model\":\"openai/gpt-4.1-nano\",\"messages\":[{\"role\":\"user\",\"content\":\"Reply with exactly $QA_CACHE_REPLY\"}],\"max_tokens\":16}"
+sed -n '1,20p' "$HEADERS_FILE"
+sed -n '1,20p' "$BODY_FILE"
+jq -e --arg reply "$QA_CACHE_REPLY" '.choices[0].message.content == $reply' "$BODY_FILE" >/dev/null
+grep -Eiq '^X-Cache: HIT \(exact\)' "$HEADERS_FILE"
```
### S73 Cache overview filtered by user path
@@ -931,9 +1032,12 @@ Checks cache analytics after the exact-cache hit using the same tracked user pat
```bash
sleep 6
-curl -sS "$AUTH_BASE_URL/admin/api/v1/cache/overview?days=1&user_path=$QA_CACHE_USER_PATH" \
+CACHE_OVERVIEW_JSON_FILE="$QA_RUN_DIR/s73.cache-overview.json"
+curl -fsS "$AUTH_BASE_URL/admin/api/v1/cache/overview?days=1&user_path=$QA_CACHE_USER_PATH" \
-H "$ADMIN_AUTH_HEADER" \
- | jq '.'
+ > "$CACHE_OVERVIEW_JSON_FILE"
+jq '.' "$CACHE_OVERVIEW_JSON_FILE"
+jq -e '.summary.total_hits >= 1 and .summary.exact_hits >= 1' "$CACHE_OVERVIEW_JSON_FILE" >/dev/null
```
### S74 Cached usage log filtered by user path
@@ -941,9 +1045,15 @@ curl -sS "$AUTH_BASE_URL/admin/api/v1/cache/overview?days=1&user_path=$QA_CACHE_
Reads cached-only usage entries for the same exact-hit request path.
```bash
-curl -sS "$AUTH_BASE_URL/admin/api/v1/usage/log?days=1&user_path=$QA_CACHE_USER_PATH&cache_mode=cached&limit=5" \
+CACHED_USAGE_JSON_FILE="$QA_RUN_DIR/s74.cached-usage.json"
+curl -fsS "$AUTH_BASE_URL/admin/api/v1/usage/log?days=1&user_path=$QA_CACHE_USER_PATH&cache_mode=cached&limit=5" \
-H "$ADMIN_AUTH_HEADER" \
- | jq '{total,entries:(.entries|map({request_id,cache_type,model,provider,endpoint,user_path,total_tokens}))}'
+ > "$CACHED_USAGE_JSON_FILE"
+jq '{total,entries:(.entries|map({request_id,cache_type,model,provider,endpoint,user_path,total_tokens}))}' "$CACHED_USAGE_JSON_FILE"
+jq -e --arg request_id "$QA_CACHE_REQ2" '
+ .total >= 1
+ and any(.entries[]?; .request_id == $request_id and .cache_type == "exact")
+ ' "$CACHED_USAGE_JSON_FILE" >/dev/null
```
### S75 Invalid managed API key user path (negative)
@@ -951,11 +1061,16 @@ curl -sS "$AUTH_BASE_URL/admin/api/v1/usage/log?days=1&user_path=$QA_CACHE_USER_
Verifies user-path validation for managed API key creation.
```bash
-curl -sS -i -X POST "$AUTH_BASE_URL/admin/api/v1/auth-keys" \
+HEADERS_FILE=$(mktemp "$QA_RUN_DIR/s75.headers.XXXXXX")
+BODY_FILE=$(mktemp "$QA_RUN_DIR/s75.body.XXXXXX")
+curl -sS -D "$HEADERS_FILE" -o "$BODY_FILE" -X POST "$AUTH_BASE_URL/admin/api/v1/auth-keys" \
-H "$ADMIN_AUTH_HEADER" \
-H 'Content-Type: application/json' \
- -d '{"name":"qa-invalid-user-path","user_path":"/team/../alpha"}' \
- | sed -n '1,20p'
+ -d '{"name":"qa-invalid-user-path","user_path":"/team/../alpha"}'
+sed -n '1,20p' "$HEADERS_FILE"
+sed -n '1,20p' "$BODY_FILE"
+grep -Eiq '^HTTP/.* 400 ' "$HEADERS_FILE"
+jq -e '.error.type == "invalid_request_error" and (.error.message | test("invalid user_path"))' "$BODY_FILE" >/dev/null
```
### S76 Invalid workflow scope user path (negative)
@@ -963,11 +1078,16 @@ curl -sS -i -X POST "$AUTH_BASE_URL/admin/api/v1/auth-keys" \
Verifies user-path validation for workflow creation.
```bash
-curl -sS -i -X POST "$AUTH_BASE_URL/admin/api/v1/execution-plans" \
+HEADERS_FILE=$(mktemp "$QA_RUN_DIR/s76.headers.XXXXXX")
+BODY_FILE=$(mktemp "$QA_RUN_DIR/s76.body.XXXXXX")
+curl -sS -D "$HEADERS_FILE" -o "$BODY_FILE" -X POST "$AUTH_BASE_URL/admin/api/v1/execution-plans" \
-H "$ADMIN_AUTH_HEADER" \
-H 'Content-Type: application/json' \
- -d '{"scope_provider":"openai","scope_model":"gpt-4.1-nano","scope_user_path":"/team/../alpha","name":"qa-invalid-workflow-path","plan_payload":{"schema_version":1,"features":{"cache":true,"audit":true,"usage":true,"guardrails":false},"guardrails":[]}}' \
- | sed -n '1,24p'
+ -d '{"scope_provider":"openai","scope_model":"gpt-4.1-nano","scope_user_path":"/team/../alpha","name":"qa-invalid-workflow-path","plan_payload":{"schema_version":1,"features":{"cache":true,"audit":true,"usage":true,"guardrails":false},"guardrails":[]}}'
+sed -n '1,24p' "$HEADERS_FILE"
+sed -n '1,24p' "$BODY_FILE"
+grep -Eiq '^HTTP/.* 400 ' "$HEADERS_FILE"
+jq -e '.error.type == "invalid_request_error" and (.error.message | test("invalid scope_user_path"))' "$BODY_FILE" >/dev/null
```
## 13. Authenticated cleanup
@@ -977,11 +1097,19 @@ curl -sS -i -X POST "$AUTH_BASE_URL/admin/api/v1/execution-plans" \
Deactivates the managed key created for the auth-enabled release run.
```bash
-AUTH_KEY_ID=$(curl -sS "$AUTH_BASE_URL/admin/api/v1/auth-keys" \
+AUTH_KEYS_JSON_FILE="$QA_RUN_DIR/s77.auth-keys.json"
+curl -fsS "$AUTH_BASE_URL/admin/api/v1/auth-keys" \
-H "$ADMIN_AUTH_HEADER" \
- | jq -r ".[] | select(.name==\"$QA_AUTH_KEY_NAME\") | .id")
-curl -sS -i -X POST "$AUTH_BASE_URL/admin/api/v1/auth-keys/$AUTH_KEY_ID/deactivate" \
+ > "$AUTH_KEYS_JSON_FILE"
+if ! AUTH_KEY_ID=$(jq -er --arg name "$QA_AUTH_KEY_NAME" '.[] | select(.name == $name) | .id' "$AUTH_KEYS_JSON_FILE"); then
+ echo "error: managed API key id not found for $QA_AUTH_KEY_NAME" >&2
+ exit 1
+fi
+HEADERS_FILE=$(mktemp "$QA_RUN_DIR/s77.headers.XXXXXX")
+curl -sS -D "$HEADERS_FILE" -o /dev/null -X POST "$AUTH_BASE_URL/admin/api/v1/auth-keys/$AUTH_KEY_ID/deactivate" \
-H "$ADMIN_AUTH_HEADER"
+sed -n '1,20p' "$HEADERS_FILE"
+grep -Eiq '^HTTP/.* 204 ' "$HEADERS_FILE"
```
### S78 Deactivated managed API key is rejected
@@ -989,13 +1117,19 @@ curl -sS -i -X POST "$AUTH_BASE_URL/admin/api/v1/auth-keys/$AUTH_KEY_ID/deactiva
Confirms that the same managed key can no longer authenticate requests.
```bash
-API_KEY=$(cat "$QA_AUTH_KEY_VALUE_FILE")
-curl -sS -i "$AUTH_BASE_URL/v1/chat/completions" \
+require_release_artifact "$QA_AUTH_KEY_VALUE_FILE"
+API_KEY=$(<"$QA_AUTH_KEY_VALUE_FILE")
+HEADERS_FILE=$(mktemp "$QA_RUN_DIR/s78.headers.XXXXXX")
+BODY_FILE=$(mktemp "$QA_RUN_DIR/s78.body.XXXXXX")
+curl -sS -D "$HEADERS_FILE" -o "$BODY_FILE" "$AUTH_BASE_URL/v1/chat/completions" \
-H "Authorization: Bearer $API_KEY" \
-H 'Content-Type: application/json' \
-H "X-Request-ID: $QA_DEACTIVATED_REQ" \
- -d '{"model":"openai/gpt-4.1-nano","messages":[{"role":"user","content":"Reply with exactly QA_AUTH_DEACTIVATED"}],"max_tokens":16}' \
- | sed -n '1,20p'
+ -d '{"model":"openai/gpt-4.1-nano","messages":[{"role":"user","content":"Reply with exactly QA_AUTH_DEACTIVATED"}],"max_tokens":16}'
+sed -n '1,20p' "$HEADERS_FILE"
+sed -n '1,20p' "$BODY_FILE"
+grep -Eiq '^HTTP/.* 401 ' "$HEADERS_FILE"
+jq -e '.error.type == "authentication_error"' "$BODY_FILE" >/dev/null
```
### S79 Deactivate scoped workflow
@@ -1003,8 +1137,12 @@ curl -sS -i "$AUTH_BASE_URL/v1/chat/completions" \
Deactivates the workflow created for the scoped managed-key release run.
```bash
-WORKFLOW_ID=$(cat "$QA_WORKFLOW_ID_FILE")
-curl -sS -i -X POST "$AUTH_BASE_URL/admin/api/v1/execution-plans/$WORKFLOW_ID/deactivate" \
+require_release_artifact "$QA_WORKFLOW_ID_FILE"
+WORKFLOW_ID=$(<"$QA_WORKFLOW_ID_FILE")
+HEADERS_FILE=$(mktemp "$QA_RUN_DIR/s79.headers.XXXXXX")
+curl -sS -D "$HEADERS_FILE" -o /dev/null -X POST "$AUTH_BASE_URL/admin/api/v1/execution-plans/$WORKFLOW_ID/deactivate" \
-H "$ADMIN_AUTH_HEADER"
+sed -n '1,20p' "$HEADERS_FILE"
+grep -Eiq '^HTTP/.* 204 ' "$HEADERS_FILE"
rm -f "$QA_AUTH_KEY_JSON" "$QA_AUTH_KEY_VALUE_FILE" "$QA_WORKFLOW_JSON" "$QA_WORKFLOW_ID_FILE"
```
diff --git a/tests/e2e/run-release-e2e.sh b/tests/e2e/run-release-e2e.sh
index 1c75f25bb..c6b4a5891 100755
--- a/tests/e2e/run-release-e2e.sh
+++ b/tests/e2e/run-release-e2e.sh
@@ -364,6 +364,7 @@ for index in "${SELECTED_INDEXES[@]}"; do
{
printf '#!/usr/bin/env bash\n'
printf 'set -euo pipefail\n'
+ printf 'shopt -s inherit_errexit 2>/dev/null || true\n'
printf 'cd %q\n' "$REPO_ROOT"
printf 'export QA_SUFFIX=%q\n' "$QA_SUFFIX"
printf 'export QA_RUN_DIR=%q\n' "$OUTPUT_DIR"
diff --git a/tests/integration/dbassert/auditlog.go b/tests/integration/dbassert/auditlog.go
index 7c637ce00..47109d644 100644
--- a/tests/integration/dbassert/auditlog.go
+++ b/tests/integration/dbassert/auditlog.go
@@ -45,7 +45,7 @@ func QueryAuditLogsByRequestID(t *testing.T, pool *pgxpool.Pool, requestID strin
defer cancel()
query := `
- SELECT id, timestamp, duration_ns, model, provider, status_code,
+ SELECT id, timestamp, duration_ns, requested_model, provider, status_code,
request_id, auth_key_id, client_ip, method, path, user_path, stream, error_type, data
FROM audit_logs
WHERE request_id = $1
@@ -141,7 +141,9 @@ func bsonToAuditLogEntry(t *testing.T, doc bson.M) AuditLogEntry {
} else if v, ok := doc["duration_ns"].(int32); ok {
entry.DurationNs = int64(v)
}
- if v, ok := doc["model"].(string); ok {
+ if v, ok := doc["requested_model"].(string); ok {
+ entry.Model = v
+ } else if v, ok := doc["model"].(string); ok {
entry.Model = v
}
if v, ok := doc["provider"].(string); ok {