Skip to main content
A lightweight liveness endpoint that confirms the gateway process is running and able to handle HTTP requests.

Endpoint

GET /health
No authentication is required.

Response

Returns 200 OK with a fixed JSON body:
{"status":"ok"}
status
string
required
Always "ok". This value is a static string written directly by the handler — no dynamic checks are performed.
This endpoint checks only that the gateway process is alive and its HTTP server is accepting connections. It does not verify that upstream models (drafter or heavyweight) are reachable, that Qdrant or Redis are healthy, or that any dependent service is available.

Use cases

Docker health check

Add a HEALTHCHECK instruction to your Dockerfile or docker-compose.yml to let Docker restart the container if the gateway becomes unresponsive.
HEALTHCHECK --interval=30s --timeout=5s --start-period=10s --retries=3 \
  CMD wget -qO- http://localhost:8080/health || exit 1

Kubernetes liveness probe

Use /health as a livenessProbe to let Kubernetes restart pods that are no longer responding.
livenessProbe:
  httpGet:
    path: /health
    port: 8080
  initialDelaySeconds: 10
  periodSeconds: 30
  timeoutSeconds: 5
  failureThreshold: 3

Load balancer health check

Most load balancers can poll /health to determine whether to include an instance in the rotation. Configure the check to expect HTTP 200 and the response body {"status":"ok"}.

Example

curl http://localhost:8080/health
Response:
{"status":"ok"}

Build docs developers (and LLMs) love