Add /status endpoint for k8s or other probe systems

[aimxhaisse]

Context: Kubernetes has a system of probes, which regularly pings a service with HTTP queries to know if it's healthy or not. If the service fails to reply it will automatically restart it / reschedule it, potentially on another machine.

This is usually implemented exporting a dummy /status endpoint which returns 200 on one of the services exported by the application, as a starting point this likely could be part of the metrics module.

[aimxhaisse]

I'll have a PR for it soon.

[aimxhaisse]

#94

[aimxhaisse]

Alternatively/likely longer term, we can introduce a cb_probe module which exports:

• /ready: once the module has loaded its config and is ready
• /status: always return a 200

This would help a lot with k8s, where we could configure it to wait for /ready to return a 200 then expose the service to your stack/delete previous deployments/and rely on /status. If there is a slower startup step at some point (for instance to load mapping of lido keys or similar), ready would return a 500 until properly loaded. That way upgrades would have 0 downtime even if there is a block proposal right in the middle, k8s will route to the previous instance until the new one is ready.

[aimxhaisse]

Marking as done for now, I'll follow up with a dedicated cb_probe module once the k8s initial draft works.