fix(core): Always serve /healthz on task broker regardless of N8N_ENDPOINT_HEALTH

[everettbu]

Mirror of n8n-io/n8n#25959
Original author: horiyee

---

Summary

This is a regression introduced by #25729, which made the health endpoint configurable across all servers including the task broker. The task-runner-launcher binary, however, hardcodes /healthz for the broker readiness check. When users set N8N_ENDPOINT_HEALTH=health — required on platforms such as Google Cloud Run that reserve /healthz at the ingress — the broker began serving health at /health while the launcher kept checking /healthz, receiving 404 and waiting forever. Code nodes would time out after 60 seconds.

The task broker runs on an internal port (5679) that is never publicly exposed, so N8N_ENDPOINT_HEALTH should not affect it. This fix hardcodes /healthz on the task broker, decoupling it from the external health endpoint configuration.

Root cause:

• Broker uses globalConfig.endpoints.health: task-broker-server.ts#L173–L174
• Launcher hardcodes /healthz: check_until_broker_ready.go#L12

Closes #25958

Testing

Added a unit test to task-broker-server.test.ts that verifies /healthz returns 200 even when N8N_ENDPOINT_HEALTH is overridden (e.g. endpoints: { health: '/health' } in GlobalConfig).

Review / Merge checklist

• PR title and summary are descriptive. (conventions)
• Docs updated or follow-up ticket created.
• Tests included.
• PR Labeled with release/backport (if the PR is an urgent fix that needs to be backported)

[everettbu]

Upstream PR was closed or merged. Code is synced via branch mirror.