rolling update for outdated pods

• All containers of selected pods get checked against their remote repositories.
• Improved logging.

rolling updates to the rescue!

When encountering a pod running on an outdated image, the k8s-auto-updater now tries to get to the corresponding deployment, then patches this environment, thus triggering a new replicaset and newly started shiny pods. So no downtime any longer fingerscrossed...

k8s-auto-updater gets a little bit more cautious!

out-of-the-box, only pods with the label auto-update=enabled get selected for further inspection
(see the description section).
You can change that selector by setting a podSelector.
It is possible to further narrow down the set of images to check by defining an imageRegExp.
And you can limit the number of deleted pods per image with nrPodsToDelete, thus giving your high-available-cluster the chance to restart pods with the newly pulled images.

See also the updated config section and the notes, too.

This is a pre-release, because per definition the k8s-auto-updater will never be production-ready.