Deploys routinely ship code but not everything the code needs: the site stays "up" while every new route 500s because a table or seed row is missing. This loop closes the merge-to-live gap end to end. Each iteration checks the next parity layer — the deploy pipeline succeeded for YOUR commit, the platform finished rolling it out, production serves the merged version, each new or changed route returns its expected status, and schema migrations plus data/content updates actually landed — and fixes the first broken layer it finds. Distinct from deploy-verification-loop (health and smoke endpoints): this verifies commit, schema, and content parity, not just liveness.
Loop
- Goal
- production fully reflects the merged commit — pipeline green, new version serving, every new or changed route returning its expected status, schema and data applied
- Exit when
- the latest main pipeline concluded success and every new or changed route on production returns its expected status code
Check between iterations
gh run list --branch main --limit 1 --json conclusion -q '.[0].conclusion' && curl -s -o /dev/null -w '%{http_code}\n' https://<your-app>/<new-route>
$
Kickoff prompt
Start the "Merge to Live Parity" loop.
Goal: production fully reflects the merged commit — pipeline green, new version serving, every new or changed route returning its expected status, schema and data applied
Max iterations: 6
Between iterations run: gh run list --branch main --limit 1 --json conclusion -q '.[0].conclusion' && curl -s -o /dev/null -w '%{http_code}\n' https://<your-app>/<new-route>
Exit when: the latest main pipeline concluded success and every new or changed route on production returns its expected status code
Step 1: List what the merge changed (routes, schema, content), then check the deploy pipeline for the merged commit and probe the first new route on production.
›