E2E auth setup timeout — pre-existing infra issue #112
Reference in New Issue
Block a user
Delete Branch "%!s()"
Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?
Bug: E2E auth setup timeout
Symptom: E2E run fails at auth.setup.ts:49 with:
Reproduction: Run E2E on PR#110 (
ci/e2e-concurrency-guard) — run #23500542756Context: This is the SAME failure mode observed in PR#109 (run #23499990163), which was attributed to concurrent runs. However, PR#110's E2E run failed even though no other PR E2E was running concurrently at the time (the concurrency guard was already in the PR branch).
Likely causes to investigate:
headlamp-e2einstance inprivilegedescalation-devmay be in a degraded state from prior runse2e/auth.setup.ts:49may have a race condition — the OIDC flow may not always show the "use a token" buttonscripts/deploy-e2e-headlamp.shdoes not clean up prior deployment before applying new oneImpact: Blocks PR approval and future PRs despite the concurrency fix being correct.
Paperclip: PRI-822
E2E is now passing on PR#110 (run #23500856092). The auth timeout was a transient issue with the shared headlamp-e2e instance. Closing this as the concurrency fix is working correctly.
The underlying cause (shared E2E instance state) may still be worth investigating separately, but it is not blocking PRs at this time.
E2E is now passing. Transient infrastructure issue, not a code bug.