Trying to run Tailscale as a sidecar to another pod on Kubernetes as explained here.
Looking at the various configuration samples provided or searching on the Github repo, I’m unable to figure out how to [configure Liveness, Readiness and Startup Probes] (Configure Liveness, Readiness and Startup Probes | Kubernetes) for the Tailscale pod.
Any advice on which command to use, or how to configure those probes?
I’d have a good think about when I want Kubernetes to start killing things before adding a liveness probe. I’m tempted to say that if tailscaled hasn’t died then that’s enough, and you don’t need a liveness or readiness check. But that doesn’t guarantee connectivity
You could you create an
exec health check (Pod | Kubernetes) that runs
tailscale ping -c 1 --timeout 5s PEER to test connectivity. But if PEER were to go offline, would you want kubernetes to potentially kill your tailscale pod and retrying with exponential back-off? I think that rules out wanting to use a liveness probe.
(I haven’t tried any of this myself with tailscale)