Readiness and Liveliness Probes

[DefinitelyNotAGoat]

My team and I really love this indexer and appreciate the hard work of the Baking Bad team. One thing that I think might be mutually beneficial to everyone is to add readiness and liveliness probes in the synchronizer/API so that users can create custom alerting and handling Kubernetes logic in the case of error.

[Groxan]

Hey, we've finally implemented health checks in v1.6.0 that can be used for liveness probes.

The Tzkt.Api implements health checks via API endpoint. Settings:

"HealthChecks": {
  "Enabled": true,
  "Endpoint": "/health"
}

After this you can use /health (or whatever you specified in the settings) endpoint. If it returns HTTP 200, then everything's ok.

The Tzkt.Sync implements health checks by creating and touching the file on the disc. Settings:

"HealthChecks": {
  "Enabled": true,
  "Delay": 10,
  "Period": 15,
  "FilePath": "/tmp/sync.health"
}

After Delay seconds you can check if the file at the specified path was modified within the last Period seconds. If yes, then everything's ok. For k8s it should look like this:

livenessProbe:
  exec:
    command:
    - find
    - /tmp/sync.health
    - -mmin
    - '-15'
  initialDelaySeconds: 10
  periodSeconds: 15