Show container pull errors in the github output

[murphpdx]

When a workflow attempts to use an image that does not exist or that the workflow does not have access to it prints out a generic error "Executing the custom container implementation failed. Please contact your self hosted runner administrator." This error is not descriptive enough for users without additional logs from the runners.

This PR updates the error messages in these cases. If the container failed due to a container error it outputs the error.

[remidebette]

+1 — we hit this. Users running into private-registry-403 or non-existent-tag see only:

##[error]Error: pod failed to come online with error: <timeout>
##[error]Executing the custom container implementation failed. Please contact your self hosted runner administrator.

The actual containerStatuses[].state.waiting.{reason,message} (carrying ImagePullBackOff and the upstream daemon's 403 / manifest-not-found body) is invisible until an operator runs kubectl describe against a pod that's usually already been pruned. This PR is the right shape and would close that gap.

(For pod-events-driven failures like FailedScheduling: Too many pods — a different K8s resource — I'm filing a separate issue: a sibling enrichment in waitForPodPhases's timeout path. Linked once filed.)

[remidebette]

Filed the sibling Events-reading enrichment as #366.

[pkking]

Would be great to have this improvement , the info is a key to help user self-solve the failed workflow when provided a wrong image @nikola-jokic please