Add a way to add readiness/liveness probes to workspace containers

[amisevsk]

Description

The DevWorkspace Operator implements a rudimentary readiness probe (curling the mainUrl for a workspace) in order to verify that a workspace is actually running before marking it as ready.

However, this is not sufficient to detect issues with the workspace after it has been started; the mainUrl may stop responding at some point in the future, and if the DevWorkspace Operator doesn't happen to reconcile the DevWorkspace, it will not detect this issue.

Instead, we consider adding a way to define readiness/liveness probes for container components. Such a probe could continue to check the main URL for the workspace and kill the pod if it becomes unhealthy.

This is currently do-able in a somewhat roundabout way by using containerOverrides, so it may also make sense to continue using that as the supported approach.

Additional context