Add some way of defining readiness/liveness checks for DevWorkspaces

[amisevsk]

Description

Currently, the DevWorkspace Operator will do a rudimentary liveness check on workspaces by making a http get request at the workspace's <mainUrl>/healthz. However, in order to not break applications that do not implement /healthz, it has to treat 4xx staus codes as successes, meaning the health check doesn't do a whole lot for most workloads.

Instead, we should define some way of specifying actual readiness/liveness probes for workspaces, allowing e.g. the Code editor running in the pod to prevent the workspace entering a running state until it is ready to serve requests.

It should be possible to implement a workaround for this by using pod-overrides and container-overrides, but perhaps a more formal attribute or similar would be useful here.

Additional context