Health & Readiness
Use /healthz and /readyz correctly.
| Endpoint | Success | Failure and meaning |
|---|---|---|
GET /healthz | 200 {"status":"ok"} | Proves only that the HTTP process responds |
GET /readyz | Default 200 {"ready":true}, or the provider's readiness body | Returns 503 with the provider body when not ready |
Both routes are always registered and have no built-in authentication. Use /healthz for process restart probes and /readyz for load-balancer admission. Readiness still does not prove spare capacity, end-to-end delivery, backup recoverability, or the safety of a high-risk change.
curl --fail http://127.0.0.1:5001/healthz
curl --fail http://127.0.0.1:5001/readyzBoth remain observable during restore maintenance. Retain the raw /readyz failure body; never restore product traffic because /healthz returned 200.
GET
/healthzReturns 200 while the HTTP process can serve requests. It does not prove cluster readiness.
Response Body
application/json
{ "status": "ok"}GET
/readyzReturns 200 only when this node can accept product traffic; 503 retains the current reason.
Response Body
application/json
application/json
{ "ready": true}