Healthchecks don't consider tunnel status

[james-callahan]

I saw in c5af83a that a /healthcheck endpoint was added to the metrics server.
However, it doesn't seem to take into account tunnel status: it returns OK even when the tunnel is not up.

Could you add a healthcheck endpoint that takes into consideration if the tunnel is actually up?

It should be a in a form suitable for a kubernetes health check.

[HofmannZ]

Any update on this?

[adamchalmers]

Hi all! We added a new healthcheck endpoint under /ready which is smarter. It will respond with HTTP 200 if cloudflared has at least one connection to the edge. It was added in 38fb0b2 and it's designed for use with k8s. Please let me know if you have any other ideas around this!

[HofmannZ]

@adamchalmers any chance you could give a quick example?

[adamchalmers]

Start cloudflared with flag --metrics localhost:3333 and then curl localhost:3333/ready to see.

If you don't specify --metrics, it'll start the server on a random port. Read the logs cloudflared outputs on startup to see which port was chosen.

[mijndert]

@adamchalmers sorry for replying to an old thread, I don't think my question warrants a new issue.

We're running cloudflared in AWS Fargate which does give me the option to of course set an entrypoint script, etc. But since curl is not included in the Dockerfile, I can't seem to figure out how to properly set up a Docker healthcheck.

Is my only option here to build my own Docker image that does include curl so I can reach the /ready endpoint and get a Docker healthcheck going?

Thanks.

[milindpatel63]

i can't seem to be able to setup a healthcheck

/bin/bash and /bin/sh not available inside docker container.

Any help?