remove backoff time from application blocking latencies

Author: daniel-sanche

packages/google-cloud-bigtable/google/cloud/bigtable/data/_metrics/data_model.py

@@ -140,7 +140,6 @@ class ActiveAttemptMetric:
     # time waiting on user to process the response, in nanoseconds
     # currently only relevant for ReadRows
     application_blocking_time_ns: int = 0
-    # backoff time is added to application_blocking_time_ns
     backoff_before_attempt_ns: int = 0
 
 

packages/google-cloud-bigtable/google/cloud/bigtable/data/_metrics/handlers/opentelemetry.py

@@ -234,8 +234,7 @@ def on_attempt_complete(
         )
         self.otel.throttling_latencies.record(flow_throttling, labels)
         self.otel.application_latencies.record(
-            (attempt.application_blocking_time_ns + attempt.backoff_before_attempt_ns)
-            / NS_TO_MS,
+            attempt.application_blocking_time_ns / NS_TO_MS,
             labels,
         )
         if attempt.gfe_latency_ns is not None:

packages/google-cloud-bigtable/tests/unit/data/_metrics/test_opentelemetry_handler.py

@@ -361,8 +361,7 @@ def test_on_attempt_complete_application_latencies(self):
         op = ActiveOperationMetric(op_type=OperationType.READ_ROWS)
         handler.on_attempt_complete(attempt, op)
         mock_instruments.application_latencies.record.assert_called_once_with(
-            (attempt.application_blocking_time_ns + attempt.backoff_before_attempt_ns)
-            / 1e6,
+            attempt.application_blocking_time_ns / 1e6,
             mock.ANY,
         )