How to perform health check with Docker Compose for Teleport?

[MillhioreFiriannoBiscotti]

I currently have Teleport deployed with Docker Compose.
According to the documentation here, I can use healthz and readyz endpoint to monitor the status of teleport. However, there seems to be no curl binary in the distroless container image.

How to perform health check with Docker Compose in this case?

[gordianbahri]

Add this health check to your teleport-server service in your docker-compose.yml:

services:
  teleport-server:
    ...
    healthcheck:
      test: tctl status
      interval: 2s
      timeout: 2s
      retries: 3
      start_period: 5s

If you also have teleport nodes in your docker compose file, you can add a condition to these services to only start them if your teleport server is running:

services:
  teleport-node:
    ...
    depends_on:
      teleport-server:
        condition: service_healthy

[gordianbahri]

Thanks @webvictim, I updated my previous answer.

[rts-kotori]

Seems like the following can be used:

    healthcheck:
      test: ["CMD", "tctl", "status"]
      interval: 2s
      timeout: 2s
      retries: 3
      start_period: 5s

And it does return healthy status.
However I am not sure about the behavior of the tctl status command. Will it return non-zero exit code if the health is degraded?

[webvictim]

No. You shouldn't use tctl status.

[MillhioreFiriannoBiscotti]

No. You shouldn't use tctl status.

What command should be used for healthcheck in compose then?

[gordianbahri]

What command should be used for healthcheck in compose then?

None. This is discouraged. Please refer to my previous answer, where I explain in depth why it's bad practice to monitor such a container within the same container and why it violates the separation of concerns principle.