[AuricX]

Summary

Adds a shard label to proxy_request_duration_seconds, so front-proxy latency can be split per backend shard, and removes the TODO(csams) in pkg/proxy/metrics/metrics.go.

The shard is resolved inside the latency middleware (by lookup.WithClusterResolver, on a derived context), so WithLatencyTracking can't read it from its own request. It now uses the same ShardNameHolder pattern as WithBodyTracking: it stores a holder in the request context before calling the next handler, and promhttp.WithLabelFromCtx reads the name from it after the handler returns. Requests the front-proxy serves itself (/metrics, /readyz) get shard="".

WithShardNameHolder now returns the existing holder when the context already has one. Otherwise the holder created by WithBodyTracking would shadow the latency one and the latency metric would never see the shard name. A small ShardNameHolderFrom getter is added so the metrics package can read the holder.

The TODO mentions the shard URL, but the label uses the shard name to match kcp_backend_rq_body_bytes, kcp_backend_rs_body_bytes and kcp_listener_tls_connection_error. Cardinality grows by the number of shards. The metric is ALPHA, and existing queries that aggregate by (code, method) keep working.

TestWithLatencyTracking wraps a handler in the real middleware order (WithLatencyTracking(WithBodyTracking(...))) and checks one observation lands on shard="shard-1", or on shard="" when no shard is set. It fails without this change because the metric has no shard label, and it also fails if the holder isn't reused.

What Type of PR Is This?

/kind feature

Related Issue(s)

Part of #4502

Release Notes

The front-proxy `proxy_request_duration_seconds` metric now has a `shard` label.

[kcp-ci-bot]

Hi @AuricX. Thanks for your PR.

I'm waiting for a kcp-dev member to verify that this patch is reasonable to test. If it is, they should reply with /ok-to-test on its own line. Until that is done, I will not automatically test new commits in this PR, but the usual testing commands by org members will still work. Regular contributors should join the org to skip this step.

Once the patch is verified, the new status will be reflected by the ok-to-test label.

I understand the commands that are listed here.

Details

Instructions for interacting with me using PR comments are available here. If you have questions or suggestions related to my behavior, please file an issue against the kubernetes-sigs/prow repository.

[ntnn]

/lgtm
/approve
/ok-to-test

Thank you!

[kcp-ci-bot]

LGTM label has been added.

Details

Git tree hash: ad2ae193dd616a696b9df8661ca0fff585dc7232

[AuricX]

/retest

[mjudeikis]

/retest

[mjudeikis]

This needs rebase now. Please rebase and convert to read. Moving to draft in the meantime

[AuricX]

@mjudeikis I rebased onto main, the only conflict came from pkg/proxy/lookup being merged into pkg/server/proxy so ShardNameHolderFrom lives there now and the metrics package imports it from there, tests and lint pass locally. @ntnn the push dropped your lgtm, please take another look when you have some time.

[mjudeikis]

/lgtm
/approve

[kcp-ci-bot]

LGTM label has been added.

Details

Git tree hash: df5575b8ba9ddf326b97565db22a28b5ffe11d6f

[kcp-ci-bot]

[APPROVALNOTIFIER] This PR is APPROVED

This pull-request has been approved by: mjudeikis, ntnn

The full list of commands accepted by this bot can be found here.

The pull request process is described here

Details

Needs approval from an approver in each of these files:

• OWNERS [mjudeikis,ntnn]

Approvers can indicate their approval by writing /approve in a comment
Approvers can cancel approval by writing /approve cancel in a comment