ci(deploy): probe backend /readyz so a dead backend fails the deploy #94

Merged

developer merged 2 commits from feature/deploy-backend-probe into development 2026-06-20 19:04:52 +00:00

Conversation 1 Commits 2 Files Changed 1

+12 -6

developer commented 2026-06-20 18:49:30 +00:00

Owner

Copy Link

Copy Source

What

Add a backend /readyz probe to the deploy job's health check.

Why

While fixing the dictionary seed-drift issue, a crash-looping backend twice passed the deploy green. Root cause: the post-deploy probe only checks GET / (static landing) and GET /app/ (gateway-served SPA shell) through the edge caddy — neither needs the backend, so a dead backend is invisible to it.

(Note: the deploy checkout is fine — it builds the PR's code, refs/pull/N/head; an earlier hunch that "PR deploys build base" was wrong, confirmed from the run #442 deploy log. The real gap was the probe.)

Change

Probe the landing, gateway and backend: add a third check, http://backend:8080/readyz on the scrabble-internal network, to the same retry loop; dump backend logs on failure. A backend that cannot become ready now fails the deploy loudly instead of going green.

Verification

This PR's own test-contour deploy exercises the new probe (the backend is healthy now → it should pass). One-line shell addition to an existing step; YAML structure unchanged.

## What Add a backend `/readyz` probe to the deploy job's health check. ## Why While fixing the dictionary seed-drift issue, a crash-looping backend twice passed the deploy **green**. Root cause: the post-deploy probe only checks `GET /` (static landing) and `GET /app/` (gateway-served SPA shell) through the edge caddy — **neither needs the backend**, so a dead backend is invisible to it. (Note: the deploy checkout is **fine** — it builds the PR's code, `refs/pull/N/head`; an earlier hunch that "PR deploys build base" was wrong, confirmed from the run #442 deploy log. The real gap was the probe.) ## Change `Probe the landing, gateway and backend`: add a third check, `http://backend:8080/readyz` on the `scrabble-internal` network, to the same retry loop; dump backend logs on failure. A backend that cannot become ready now **fails the deploy loudly** instead of going green. ## Verification This PR's own test-contour deploy exercises the new probe (the backend is healthy now → it should pass). One-line shell addition to an existing step; YAML structure unchanged.

developer added 1 commit 2026-06-20 18:49:31 +00:00

ci(deploy): probe backend /readyz so a dead backend fails the deploy ... 649b90024f

The post-deploy health check probed only the static landing (`GET /`) and the
gateway-served SPA shell (`GET /app/`) through the edge caddy. Neither needs the
backend, so a crash-looping backend passed the deploy green — which masked the
seed-drift crash until it was inspected by hand.

Add a third probe, `http://backend:8080/readyz` on the internal network, to the
same retry loop, and dump the backend logs on failure. A backend that cannot
become ready now fails the deploy loudly instead of going green.

developer added 1 commit 2026-06-20 18:56:30 +00:00

ci: re-trigger deploy to rebuild the contour app tier cleanly 76b62dea1f

owner approved these changes 2026-06-20 19:04:14 +00:00

developer merged commit 9824214fd7 into development 2026-06-20 19:04:52 +00:00

developer deleted branch feature/deploy-backend-probe 2026-06-20 19:04:52 +00:00

Sign in to join this conversation.

Reviewers

No Reviewers

owner

Labels

No items

No Label

Milestone

No items

No Milestone

Projects

Clear projects

No project

Assignees

Clear assignees

owner (Ilia) developer

No Assignees

2 Participants

Notifications

Subscribe

Due Date

No due date set.

Dependencies

No dependencies set.

Reference: developer/scrabble-game#94