Returns 200 as long as the HTTP process is up and able to serve requests at all — independent of whether models have finished loading.
​
string
Always "ok" when this endpoint responds at all.
200 OK
Use this for a liveness probe only (e.g. Kubernetes livenessProbe — “should this container be restarted?”). Do not gate a readinessProbe on this endpoint, since it returns 200 even while models are still loading and the process would reject real traffic. Use /v1/ready for that.