fix(pod): don't report Init:N/M for a started sidecar that is not ready - #4311

Problem

A pod with a native sidecar (an init container with restartPolicy: Always) whose readiness probe is failing shows Init:0/1 in the STATUS column, even though the sidecar has started and the main containers are running. kubectl shows the same pod as Running.

In checkInitContainerStatus, a started restartable init container only counts as done when it is also ready. Otherwise it falls through to Init:N/M. kubectl's printPod treats a started sidecar as past initialization whether or not it is ready (readiness only affects the READY count):

case isRestartableInitContainer(initContainers[container.Name]) &&
    container.Started != nil && *container.Started:
    if container.Ready {
        readyContainers++
    }
    continue

k9s's own initContainerStatus, used by PodStatus, already skips sidecars in this case, so the two code paths also disagreed.

Fix

A started restartable init container no longer blocks the phase. Readiness is still reflected in the READY column and in the pod health check, which haven't changed.

Tests

• New TestCheckPhase case restartable-started-not-ready: a started, not-ready sidecar plus a running, ready main container. It now gives Running; before the fix it gave Init:0/1.
• Updated the restart case of Test_checkInitContainerStatus, which recorded the old behaviour.

go test ./internal/render/... ./internal/view/... passes.

AI-assisted: I directed the investigation and reviewed and verified the change myself.