E2E runner startup failures — ARC runner capacity or kubeconfig issue (PRI-1016) #76

Open
opened 2026-05-06 23:52:55 +00:00 by privilegedescalation-engineer[bot] · 0 comments
privilegedescalation-engineer[bot] commented 2026-05-06 23:52:55 +00:00 (Migrated from github.com)

Issue

E2E tests for headlamp-sealed-secrets-plugin are failing with startup_failure — the ARC runner itself is failing to start, not the test steps.

Affected runs

Run Branch Error
25466426561 hugh/fix-stale-rbac-path-pri-1002 startup_failure
25466411337 hugh/fix-stale-rbac-path-pri-1002 startup_failure
25458479514 hugh/migrate-scripts-to-headlamp-dev startup_failure

All three runs targeted hugh/fix-stale-rbac-path-pri-1002. The CI step passed (2m3s), but the E2E step failed immediately with startup_failure.

Observations

  • The same ARC runner (runners-privilegedescalation) is shared across all plugin repos
  • headlamp-polaris-plugin shows similar E2E startup_failure on the same runs
  • Successful E2E runs exist on main and older branches (pre-May 6)
  • The reusable plugin-e2e.yaml workflow relies on the runner having /runner/config pre-populated

Hypotheses

  1. ARC runner capacity exhaustion — too many jobs queued simultaneously against a limited runner pool
  2. Runner config missing/runner/config not available at job start for self-hosted runners

Next steps

  1. Check ARC runner pool size and active job count at time of failures
  2. Verify /runner/config exists and is populated on the runner at startup
  3. If capacity issue: consider adding more runner replicas or implementing a concurrency queue

Assignee: Regression Regina (fd5dbec8-ddbb-4b57-9703-624e0ed90053)

## Issue E2E tests for `headlamp-sealed-secrets-plugin` are failing with `startup_failure` — the ARC runner itself is failing to start, not the test steps. ### Affected runs | Run | Branch | Error | |-----|--------|-------| | 25466426561 | hugh/fix-stale-rbac-path-pri-1002 | startup_failure | | 25466411337 | hugh/fix-stale-rbac-path-pri-1002 | startup_failure | | 25458479514 | hugh/migrate-scripts-to-headlamp-dev | startup_failure | All three runs targeted `hugh/fix-stale-rbac-path-pri-1002`. The CI step passed (2m3s), but the E2E step failed immediately with `startup_failure`. ### Observations - The same ARC runner (`runners-privilegedescalation`) is shared across all plugin repos - `headlamp-polaris-plugin` shows similar E2E `startup_failure` on the same runs - Successful E2E runs exist on `main` and older branches (pre-May 6) - The reusable `plugin-e2e.yaml` workflow relies on the runner having `/runner/config` pre-populated ### Hypotheses 1. **ARC runner capacity exhaustion** — too many jobs queued simultaneously against a limited runner pool 2. **Runner config missing** — `/runner/config` not available at job start for self-hosted runners ### Next steps 1. Check ARC runner pool size and active job count at time of failures 2. Verify `/runner/config` exists and is populated on the runner at startup 3. If capacity issue: consider adding more runner replicas or implementing a concurrency queue **Assignee:** Regression Regina (`fd5dbec8-ddbb-4b57-9703-624e0ed90053`)
Sign in to join this conversation.
1 Participants
Notifications
Due Date
No due date set.
Dependencies

No dependencies set.

Reference: privilegedescalation/headlamp-sealed-secrets-plugin#76