flake: Test_localSub_init/CrossSubjectListenerIsolation

[flake-investigator]

CI Run Link: https://github.com/coder/coder/actions/runs/26576090091
Failing Job: https://github.com/coder/coder/actions/runs/26576090091/job/78296339738 (test-go-pg-17)
Failure time: 2026-05-28T13:08:36Z (same window as Slack alert)

Commit Info:

• SHA: 12520ee964270e561e0e1b75b765a4194c159c4a
• Author: Danny Kopping
• Commit: coder/coder@12520ee

Failure evidence:

=== FAIL: coderd/x/nats Test_localSub_init/CrossSubjectListenerIsolation (25.04s)
    pubsub_internal_test.go:321:
        Error Trace: /home/runner/work/coder/coder/coderd/x/nats/pubsub_internal_test.go:321
        Error:       Condition never satisfied
        Test:        Test_localSub_init/CrossSubjectListenerIsolation

Error analysis:

• 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.

Precise assignment analysis:

• Intended blame command:
  • git blame -L 316,333 coderd/x/nats/pubsub_internal_test.go
• Recent file history:
  • git log --oneline -10 --follow coderd/x/nats/pubsub_internal_test.go
  • Most recent substantive change: f6f284e “feat: add initial NATS implementation (#25602)” by Jon Ayers.
• Jon Ayers corresponds to GitHub user @sreya per team mapping.

Assignment:

• Best candidate: @sreya (owner of initial NATS implementation; please reassign if ownership differs).

Related issues search (coder/internal):

• "Test_localSub_init"
• "CrossSubjectListenerIsolation"
• "pubsub_internal_test.go" / "coderd/x/nats"
• "Condition never satisfied"
  No matches found.

Reproduction (best effort):

go test ./coderd/x/nats -run Test_localSub_init/CrossSubjectListenerIsolation -count=1

[flake-investigator]

New occurrence in main CI:

• CI run: https://github.com/coder/coder/actions/runs/26582325717
• Failing job: https://github.com/coder/coder/actions/runs/26582325717/job/78320555741 (test-go-pg ubuntu-latest)
• Failure time: 2026-05-28T15:07:26Z (same day as Slack alert)

Commit info:

• SHA: 6df1536256475730d2fdba72083d3142c3089a25
• Author: Cian Johnston
• Commit: coder/coder@6df1536

Failure evidence:

=== FAIL: coderd/x/nats Test_localSub_init/CrossSubjectListenerIsolation (25.06s)
    pubsub_internal_test.go:321:
        Error Trace: /home/runner/work/coder/coder/coderd/x/nats/pubsub_internal_test.go:321
        Error:       Condition never satisfied
        Test:        Test_localSub_init/CrossSubjectListenerIsolation

No data race / panic / OOM indicators observed in the job logs around the failure.

[flake-investigator]

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:

=== FAIL: coderd/x/nats Test_localSub_init/CrossSubjectListenerIsolation (25.05s)
    pubsub_internal_test.go:321:
        Error: Condition never satisfied

Notes:

• No panic/OOM/data race indicators found in the retrieved logs.

[flake-investigator]

New occurrence:

• CI Run: https://github.com/coder/coder/actions/runs/26714105970
• Failing Job: https://github.com/coder/coder/actions/runs/26714105970/job/78729525296 (test-go-pg (ubuntu-latest))
• Failure time: 2026-05-31T13:43:20Z (same window as Slack alert)

Commit Info:

• SHA: 9c111a2be2ed87490083a1224e86f04cb7bcc6ba
• Author: Dean Sheather
• Commit: coder/coder@9c111a2

Failure evidence:

=== FAIL: coderd/x/nats Test_localSub_init/CrossSubjectListenerIsolation (25.05s)
    pubsub_internal_test.go:316:
        Error Trace: /home/runner/work/coder/coder/coderd/x/nats/pubsub_internal_test.go:316
        Error:       Condition never satisfied
        Test:        Test_localSub_init/CrossSubjectListenerIsolation

Notes:

• Same failure signature as before (fast subscriber did not reach total within testutil.WaitLong).
• No data race warnings, panic traces, or OOM indicators found in the job logs around the failure.

[flake-investigator]

New occurrence:

• CI Run: https://github.com/coder/coder/actions/runs/26747280992
• Job: https://github.com/coder/coder/actions/runs/26747280992/job/78825984726 (test-go-pg-17)
• Failure time: 2026-06-01T09:49:59Z
• Commit: 8b7e040105c10e9a5bad5689da607c1f32f9c9e7 (Mathias Fredriksson) coder/coder@8b7e040

Failure evidence:

=== FAIL: coderd/x/nats Test_localSub_init/CrossSubjectListenerIsolation (25.04s)
    pubsub_internal_test.go:316:
        Error Trace: /home/runner/work/coder/coder/coderd/x/nats/pubsub_internal_test.go:316
        Error:       Condition never satisfied
        Test:        Test_localSub_init/CrossSubjectListenerIsolation

Notes:

• 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.