Accessing GPTS liveness & readiness endpoints

Liveness and readiness probe are the way that GPTS uses to inform other components about its health. Those are primarily used by Kubernetes, but can be also checked manually.

No content returned

Both endpoints are optimized to return no content and use only status code to indicate whether the service is running (200 OK) or not (503 Service Unavailable).

Liveness endpoint

Liveness endpoint indicates whether application itself (not the exposed service) is up and running.

# Change ${GPTS} to proper GPTS healthchecks URL (usually exposed on port 8081)
curl ${GPTS}/live

Example command execution & output

curl http://localhost:8081/live -w "${http_code}"

{}
200

Readiness endpoint

Readiness endpoint indicates whether both application and the exposed service are up and running.

# Change ${GPTS} to proper GPTS healthchecks URL (usually exposed on port 8081)
curl ${GPTS}/ready

Example command execution & output

curl http://localhost:8081/ready -w "${http_code}"

{}
200

Possible statuses

Application status	Service status	Liveness probe status	Readiness probe status
not ready	not ready	503 Service Unavailable	503 Service Unavailable
ready	not ready	200 OK	503 Service Unavailable
ready	ready	200 OK	200 OK

Prohibited combination

It is impossible for application to be not ready and service to be ready, so such configuration was omitted in the table above.

---

Last update: 2022-03-06