You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
The failing assertion is the require.Eventually that expects fastCount >= total in Test_localSub_init/CrossSubjectListenerIsolation.
This indicates the fast subscriber did not receive all expected messages within testutil.WaitLong when a slow subscriber is blocked, despite the test expecting isolation.
Root cause classification:
Likely flaky test / timing/backpressure in the NATS local pubsub implementation when a slow subscriber blocks and a fast subscriber should still drain.
Data race / panic / OOM checks:
No WARNING: DATA RACE, panic:, or OOM indicators observed in the job logs around the failure.
No panic, OOM, or data race warnings observed in this run’s logs around the failure.
Most recent substantive change in this test came from 5cdc9e2 “feat: add nats cluster peer support (#25632)” (Jon Ayers), which refactored this test and nearby helpers.
=== FAIL: coderd/x/nats Test_localSub_init/CrossSubjectListenerIsolation (25.06s)
pubsub_internal_test.go:317:
Error: Condition never satisfied
The failing line is the require.Eventually waiting for fastCount >= total (line 317 in coderd/x/nats/pubsub_internal_test.go).
Notes:
No data race, panic, or OOM indicators observed in this job’s logs.
File history still points to 5cdc9e28a984 (“feat: add nats cluster peer support”) as the most recent substantive change to this test, so current assignment still looks correct.
Activity
flake-investigator commentedon May 28, 2026
New occurrence in main CI:
Commit info:
Failure evidence:
No data race / panic / OOM indicators observed in the job logs around the failure.
flake-investigator commentedon May 28, 2026
New occurrence on main:
CI Run Link: https://github.com/coder/coder/actions/runs/26586696037
Failing Job: https://github.com/coder/coder/actions/runs/26586696037/job/78334548165 (test-go-pg-17, run_attempt 1)
Commit: coder/coder@094fe97 (DevCats) - "chore(aibridge): add AWS PRM user-agent attribution for Bedrock calls (#25221)"
Failure evidence:
Notes:
flake-investigator commentedon May 31, 2026
New occurrence:
Commit Info:
Failure evidence:
Notes:
totalwithintestutil.WaitLong).flake-investigator commentedon Jun 1, 2026
New occurrence:
Failure evidence:
Notes:
flake-investigator commentedon Jun 2, 2026
New occurrence in main:
Commit:
Failure evidence:
Root cause signal: flaky timing/backpressure in local NATS pubsub test (fast subscriber did not satisfy eventually condition).
flake-investigator commentedon Jun 2, 2026
New occurrence on main:
Commit:
Failure evidence:
Notes:
flake-investigator commentedon Jun 3, 2026
New occurrence:
Failure evidence:
No WARNING: DATA RACE, panic, or OOM indicators observed in the job logs around the failure.
flake-investigator commentedon Jun 4, 2026
New occurrence in
test-go-pg (ubuntu-latest).CI Run: https://github.com/coder/coder/actions/runs/26942567262
Failing job: https://github.com/coder/coder/actions/runs/26942567262/job/79487446261
Failure time: 2026-06-04T09:22:37Z
Commit:
Failure evidence:
The failing line is the
require.Eventuallywaiting forfastCount >= total(line 317 incoderd/x/nats/pubsub_internal_test.go).Notes:
5cdc9e28a984(“feat: add nats cluster peer support”) as the most recent substantive change to this test, so current assignment still looks correct.fix: fix flakiness in nats localsub test (#26084)