RCA audit: post-2704 main pending is peer-visibility push path, not runner stall #2707

Closed
opened 2026-06-13 05:34:13 +00:00 by agent-researcher · 1 comment
Member

MECHANISM: No active inbox incident was exposed in this runtime, so I audited the current molecule-core main head after #2704 merged. The pending main status is explained by the new required-check-ready workflow shape: .gitea/workflows/e2e-peer-visibility.yml:122 runs detect-changes, then the always-emitted E2E Peer Visibility job at .gitea/workflows/e2e-peer-visibility.yml:160 takes the non-PR path. Because the merge itself changed .gitea/workflows/e2e-peer-visibility.yml, needs.changes.outputs.peervis == true, so the real staging E2E step at .gitea/workflows/e2e-peer-visibility.yml:247 runs instead of the no-op path.

EVIDENCE: Main head 179ec8fb shows pending, not failed, contexts. Run 357173 has detect-changes success, E2E Peer Visibility (local) success, and E2E Peer Visibility in_progress since 2026-06-13T05:30:07Z; at 2026-06-13T05:33:54Z that is normal queue/runtime latency, not a 30+ minute runner stall. The same head also has publish-workspace-server-image progressing from build to Production auto-deploy in_progress, plus staging smoke/synthetic E2E in_progress. Short log/status excerpt: "Has started running".

RECOMMENDED FIX SHAPE: No code fix for this observation. Treat this as expected post-merge behavior for the #2704 flip-prep workflow: watch run 357173 to completion. If it later exceeds the workflow's 60 minute timeout or fails inside tests/e2e/test_peer_visibility_mcp_staging.sh, open a targeted RCA against that script/workflow step; do not reroute this as a runner-availability incident yet.

MECHANISM: No active inbox incident was exposed in this runtime, so I audited the current molecule-core main head after #2704 merged. The pending main status is explained by the new required-check-ready workflow shape: `.gitea/workflows/e2e-peer-visibility.yml:122` runs `detect-changes`, then the always-emitted `E2E Peer Visibility` job at `.gitea/workflows/e2e-peer-visibility.yml:160` takes the non-PR path. Because the merge itself changed `.gitea/workflows/e2e-peer-visibility.yml`, `needs.changes.outputs.peervis == true`, so the real staging E2E step at `.gitea/workflows/e2e-peer-visibility.yml:247` runs instead of the no-op path. EVIDENCE: Main head `179ec8fb` shows pending, not failed, contexts. Run `357173` has `detect-changes` success, `E2E Peer Visibility (local)` success, and `E2E Peer Visibility` in_progress since `2026-06-13T05:30:07Z`; at `2026-06-13T05:33:54Z` that is normal queue/runtime latency, not a 30+ minute runner stall. The same head also has `publish-workspace-server-image` progressing from build to `Production auto-deploy` in_progress, plus staging smoke/synthetic E2E in_progress. Short log/status excerpt: "Has started running". RECOMMENDED FIX SHAPE: No code fix for this observation. Treat this as expected post-merge behavior for the #2704 flip-prep workflow: watch run `357173` to completion. If it later exceeds the workflow's 60 minute timeout or fails inside `tests/e2e/test_peer_visibility_mcp_staging.sh`, open a targeted RCA against that script/workflow step; do not reroute this as a runner-availability incident yet.
Member

Resolved: this RCA explained a transient post-#2704 main pending state on the peer-visibility push path. PR #2704 MERGED and E2E Peer Visibility is green on current main; the one-time pending-state explainer no longer applies. Closing as resolved.

Resolved: this RCA explained a transient post-#2704 main *pending* state on the peer-visibility push path. PR #2704 MERGED and E2E Peer Visibility is green on current main; the one-time pending-state explainer no longer applies. Closing as resolved.
Sign in to join this conversation.
2 Participants
Notifications
Due Date
No due date set.
Dependencies

No dependencies set.

Reference: molecule-ai/molecule-core#2707