[FEATURE] Fail fast in wait-deploy-merge when deploy PR checks fail

[c-vigo]

Description

The wait-deploy-merge job in the smoke-test repository-dispatch.yml workflow should detect failed CI checks on the deploy PR and exit early, rather than waiting for the full 30-minute timeout.

Problem Statement

When the deploy PR's CI checks fail (e.g. PR #40), auto-merge never fires and the PR remains in OPEN state. The current polling loop only checks for MERGED or CLOSED states, so it keeps polling for up to 1800 seconds before timing out. This wastes ~30 minutes of CI runner time and delays failure feedback to the upstream release pipeline.

Proposed Solution

Enhance the polling loop in the wait-deploy-merge job (assets/smoke-test/.github/workflows/repository-dispatch.yml) to also query the PR's check status on each iteration. Strategy:

1. After checking state, query the PR's status checks via gh pr checks "${PR_URL}" --json bucket,state (or gh pr view --json statusCheckRollup).
2. If all checks have completed and any have failed, exit immediately with an error -- auto-merge will never proceed.
3. If checks are still pending, continue polling as before.

Sketch of the additional check inside the loop:

CHECKS_JSON="$(gh pr checks "${PR_URL}" --json bucket 2>/dev/null || echo '[]')"
FAIL_COUNT="$(echo "${CHECKS_JSON}" | jq '[.[] | select(.bucket == "fail")] | length')"
PENDING_COUNT="$(echo "${CHECKS_JSON}" | jq '[.[] | select(.bucket == "pending")] | length')"
if [ "${FAIL_COUNT}" -gt 0 ] && [ "${PENDING_COUNT}" -eq 0 ]; then
  echo "ERROR: deploy PR checks failed -- auto-merge will not proceed: ${PR_URL}"
  exit 1
fi

Alternatives Considered

• Webhook-based approach: Replace polling with a workflow_run trigger on the downstream repo. More complex, requires additional workflow wiring, and does not fit the current cross-repo dispatch pattern as directly.
• Reduce timeout: Lowering timeout still wastes runner time and can create false timeouts on slow CI runs.
• Check only after N iterations: Simpler, but still delays feedback.

Additional Context

• Downstream deploy PR: #40
• Waiting job run: actions job 67774383060

Impact

• Faster failure feedback when deploy PR checks fail.
• Reduced wasted CI runtime on known-failed deploy PRs.
• No behavior change on successful merge path.

Changelog Category

Changed

Acceptance Criteria

• wait-deploy-merge exits within 2 polling cycles when deploy PR checks fail
• wait-deploy-merge still waits correctly when checks are pending
• wait-deploy-merge still succeeds when PR merges normally
• TDD compliance (see .cursor/rules/tdd.mdc)