Health Checks¶

Performing Health Checks¶

To check a Review Board server’s health, perform a HTTP GET on the /health/ endpoint on your server (for example: https://reviews.example.com/health/).

This will return an HTTP status code and JSON payload representing the health of the Review Board server.

A healthy server will send a HTTP 200 OK response with the following JSON payload:

{
    "checks": {
        "cache.default": "UP",
        "database.default": "UP"
    },
    "errors": {},
    "status": "UP"
}

A server that’s failing will send a HTTP 503 Service Unavailable response with a JSON payload looking something like:

{
    "checks": {
        "cache.default": "DOWN",
        "database.default": "DOWN"
    },
    "errors": {
        "cache.default": "Unable to communicate with the cache server",
        "database.default": "Connection timed out"
    },
    "status": "DOWN"
}

If any services are down, the main status field will be DOWN.

Docker¶

Our official Docker images are pre-configured to perform health checks.

To enable this in your own image, run:

HEALTHCHECK CMD curl -f http://127.0.0.1/health/ || exit 1

If this is being configured on an external server, make sure to use the correct URL above and configure access to the health check endpoint.

You can configure the interval between health checks, how long until the first health check, and how many failures until a server is considered unhealthy. See the Docker HEALTHCHECK documentation for instructions.

Kubernetes¶

You can configure a HTTP liveness probe in Kubernetes to check Review Board’s health. Note that you may need to configure access to the health check endpoint for your pod.

A liveness probe may look like:

apiVersion: v1
kind: Pod
metadata:
  name: reviewboard-liveness-http
  labels:
    test: liveness

spec:
  containers:
    - name: liveness
      image: k8s.gcr.io/liveness
      args:
        - /server

      livenessProbe:
        httpGet:
          path: /health/
          port: 8080

You can configure the interval between health checks, how long until the first health check, and how many failures until a server is considered unhealthy. See the Kubernetes liveness probe documentation for instructions.

Configuring Access¶

By default, health checks can only be accessed by clients on the same server, using http://127.0.0.1/health/ or http://[::1]/health/.

To allow external servers to check the health of Review Board, you will need to add its IP address to HEALTHCHECK_IPS in your site directory’s conf/settings_local.py file.

For example:

HEALTHCHECK_IPS = [
    '10.0.1.20',
]