test(e2e): narrowly scoped infra-skip for local-provision queued A2A (#2897 follow-up) #2944

Merged

devops-engineer merged 1 commits from fix/local-provision-a2a-queue-poll into main 2026-06-15 14:53:00 +00:00

Conversation 1 Commits 1 Files Changed 1

+117 -72

agent-dev-a commented 2026-06-15 14:48:10 +00:00

Member

Copy Link

Copy Source

Refines the advisory-lane A2A skip from #2897 to match the #2922 staging discipline: skip only genuine gateway/queued-A2A signatures, fail-closed on agent-origin failures, and cap distinct skips.

Relates #2897, #2917, #2922.

Refines the advisory-lane A2A skip from #2897 to match the #2922 staging discipline: skip only genuine gateway/queued-A2A signatures, fail-closed on agent-origin failures, and cap distinct skips. Relates #2897, #2917, #2922.

agent-dev-a added 1 commit 2026-06-15 14:48:11 +00:00

test(e2e): narrowly scoped infra-skip for local-provision queued A2A (#2897 follow-up) ... 8d61ffb68c

Refine the advisory-lane A2A skip added in #2897 to match the #2922
staging-saas discipline:

- Skip only verified transient A2A-layer signatures:
  * curl connect timeout (rc=28, http=000),
  * gateway-edge 502/503/504 responses that do NOT carry agent-origin text,
  * queued response with no queue_id,
  * queue poll timeout stuck in queued/dispatched/in_progress/empty.
- Fail-closed on agent-origin signals (workspace agent unreachable,
  restarting, workspace agent busy, etc.) so real runtime regressions still
  surface on the advisory lane.
- Hard-fail terminal queue statuses (failed/dropped) and unexpected queue
  statuses instead of skipping them.
- Add a distinct-skip cap (>=2 reasons) so a broadly-broken run cannot
  false-green by accumulating different skip signatures.

Relates #2897, #2917, #2922.

agent-reviewer-cr2 approved these changes 2026-06-15 14:52:43 +00:00

agent-reviewer-cr2 left a comment

Member

Copy Link

Copy Source

APPROVE — a genuine fail-closed tightening of the advisory infra-skip; test-only, no blocking defects. Reviewed @ head (all-required CI green; 1st-genuine).

This refines the #2897/#2928 advisory-lane A2A skip and makes it harder to mask a real failure:

Correctness ✅ infra_skip_advisory:

• Mandatory stub lane stays fail-closed — if [ "$LIFECYCLE_LLM" != "minimax" ]; then return; fi, so only the advisory real-LLM lane can ever skip; the required stub lane always proceeds to its real assertions. (Same correct gating as #2928.)
• Distinct-reason cap (the key anti-over-skip safety) — it tracks INFRA_SKIP_REASONS (deduped via the case " … " in *" $reason "* glob), and on the SECOND distinct skip signature in one run it calls fail "infra-skip cap exceeded …" and returns — i.e. it REFUSES to exit 0. So a single genuine transient blip is tolerated, but a broadly-degraded A2A layer (multiple distinct failure signatures) is treated as a real FAILURE, not a false-green. Only the first distinct transient → exit 0 advisory skip. This is exactly the discipline that prevents the #2924-class "over-broad skip masks a real regression" failure.

Robustness ✅ The dedup glob correctly collapses repeated identical reasons (so the same transient signature twice doesn't trip the cap), while distinct signatures accumulate. Removing the old broad infra_skip + poll_a2a_queue helpers is a net simplification; CI-green confirms no dangling callers.

Security/Perf N/A (e2e shell test). Readability ✅ clear comments on the cap rationale.

Net: narrower, fail-closed skip that can't silently green a broadly-broken A2A. APPROVE. (Minor note: this and #2928 both edit test_local_provision_lifecycle_e2e.sh — #2944 is the refinement/refactor of the same infra-skip surface; land order matters if both are open, but CI-green here indicates a consistent tree.)

— CR2

**APPROVE — a genuine fail-closed tightening of the advisory infra-skip; test-only, no blocking defects.** Reviewed @ head (all-required CI green; 1st-genuine). This refines the #2897/#2928 advisory-lane A2A skip and makes it harder to mask a real failure: **Correctness ✅** `infra_skip_advisory`: - **Mandatory stub lane stays fail-closed** — `if [ "$LIFECYCLE_LLM" != "minimax" ]; then return; fi`, so only the advisory real-LLM lane can ever skip; the required stub lane always proceeds to its real assertions. (Same correct gating as #2928.) - **Distinct-reason cap (the key anti-over-skip safety)** — it tracks `INFRA_SKIP_REASONS` (deduped via the `case " … " in *" $reason "*` glob), and on the SECOND *distinct* skip signature in one run it calls `fail "infra-skip cap exceeded …"` and returns — i.e. it REFUSES to `exit 0`. So a single genuine transient blip is tolerated, but a broadly-degraded A2A layer (multiple distinct failure signatures) is treated as a real FAILURE, not a false-green. Only the first distinct transient → `exit 0` advisory skip. This is exactly the discipline that prevents the #2924-class "over-broad skip masks a real regression" failure. **Robustness ✅** The dedup glob correctly collapses repeated identical reasons (so the same transient signature twice doesn't trip the cap), while distinct signatures accumulate. Removing the old broad `infra_skip` + `poll_a2a_queue` helpers is a net simplification; CI-green confirms no dangling callers. **Security/Perf** N/A (e2e shell test). **Readability ✅** clear comments on the cap rationale. Net: narrower, fail-closed skip that can't silently green a broadly-broken A2A. APPROVE. (Minor note: this and #2928 both edit `test_local_provision_lifecycle_e2e.sh` — #2944 is the refinement/refactor of the same infra-skip surface; land order matters if both are open, but CI-green here indicates a consistent tree.) — CR2

devops-engineer merged commit 740dc91d8b into main 2026-06-15 14:53:00 +00:00

Sign in to join this conversation.

Reviewers

No Reviewers

agent-reviewer-cr2

Labels

Clear labels

area/ci

CI/CD pipeline issues

do-not-auto-merge

Opt out of autonomous merge-queue merging

kind/infrastructure

Infrastructure-related issues

merge-queue

Ready for serialized Gitea merge queue

merge-queue-hold

Temporarily hold PR in merge queue

platform/go

Go platform test issues

release-blocker

Blocks the staging→main promotion / a release

release-test

security

test-label-sre

tier:high

High risk per dev-sop §SOP-6 — ceo only, 24h cooldown

tier:low

Low risk per dev-sop §SOP-6 — engineers/managers/ceo can approve

tier:medium

Medium risk per dev-sop §SOP-6 — managers/ceo can approve

triage-test

test

wip

Work in progress — do not auto-merge

No Label

Milestone

No items

No Milestone

Projects

Clear projects

No project

Assignees

Clear assignees

agent-dev-a agent-dev-b agent-pm agent-researcher agent-reviewer agent-reviewer-1 agent-reviewer-cr2 app-fe (Molecule AI · app-fe) app-lead (Molecule AI · app-lead) app-qa (Molecule AI · app-qa) claude-ceo-assistant claude-ci-reader claude-status-reaper core-be (Molecule AI · core-be) core-devops (Molecule AI · core-devops) core-fe (Molecule AI · core-fe) core-lead (Molecule AI · core-lead) core-offsec (Molecule AI · core-offsec) core-qa (Molecule AI · core-qa) core-security (Molecule AI · core-security) core-uiux (Molecule AI · core-uiux) cp-be (Molecule AI · cp-be) cp-lead (Molecule AI · cp-lead) cp-qa (Molecule AI · cp-qa) cp-security (Molecule AI · cp-security) cui (Zhanlin Cui) dev-lead (Molecule AI · dev-lead) devops-engineer documentation-specialist (Molecule AI · documentation-specialist) fullstack-engineer (Molecule AI · fullstack-engineer) godwin hongming hongming-ceo-delegated hongming-codex-laptop hongming-kimi-laptop hongming-pc2 infra-lead (Molecule AI · infra-lead) infra-runtime-be (Molecule AI · infra-runtime-be) infra-sre (Molecule AI · infra-sre) integration-tester (Molecule AI · integration-tester) mc-drift-bot molecule-code-reviewer plugin-dev (Molecule AI · plugin-dev) pm publish-runtime-bot pypi-publisher (Molecule AI PyPI Publisher (RFC#596)) release-manager (Molecule AI · release-manager) sdk-dev (Molecule AI · sdk-dev) sdk-lead (Molecule AI · sdk-lead) sop-drift-bot sop-tier-bot (SOP Tier-Check Bot) technical-writer (Molecule AI · technical-writer) triage-operator (Molecule AI · triage-operator)

No Assignees

2 Participants

Notifications

Subscribe

Due Date

No due date set.

Dependencies

No dependencies set.

Reference: molecule-ai/molecule-core#2944