[Feature]: Healthcheck

[mrjvs]

Checked Existing

• I have checked the repository for duplicate issues.

What feature do you want to see added?

A way to healthcheck the service.

Why do you want to have this feature?

This is useful in infrastructure automations. It allows automation tooling to determine when a service is ready for connections or when a service is in an unrecoverable state in which it needs to restart.

Any other details to share? (OPTIONAL)

Healthcheck guidelines:

• A healthcheck endpoint can be two types:
  • A HTTP endpoint on /health that responds with 200 if healthy and 500 if unhealthy
  • A GRPC API that implements the healthcheck spec: https://github.com/grpc/grpc/blob/master/doc/health-checking.md
• A healthcheck should report unhealthy if it hasn't finished initializing yet:
  • Still loading code: unhealthy
  • Loading files from file system: unhealthy
  • Connecting to a third party service: healthy
• A healthcheck should only report unhealthy after initialisation if it's truly unrecoverable:
  • Deadlocked?: unhealthy
  • Cant reach db: still healthy
  • Can't read any files from disk: unhealthy
  • Note: You may still return the status of third party services, just dont mark it as unhealthy if it fails that check.
• A good minimal healthcheck would be one that returns healthy if it's done initializing (without counting third party connections)

[jonbarrow]

Is there a typo here or am I misunderstanding something? You say both:

A healthcheck should report unhealthy if it hasn't finished initializing

and

A healthcheck should only report unhealthy if it's truly unrecoverable

Which are mutually exclusive things

[mrjvs]

you're right @jonbarrow, I updated the wording a bit. its unhealthy only if it hasnt finished initializing or if its in an unrecoverable state.