Perform a rolling upgrade when updating the CONTAINER_IMAGE_TAG to another version

[evertmulder]

Currently we update our production environment quite often. When using the orchestrator and updating the CONTAINER_IMAGE_TAG to a newer version every pod is removed first and recreated, leading to at least 2-3 min of downtime of the API and UI. It would be nice up the existing deployment would be updated triggering an rolling update.
This would be especially useful for the UI and Webservice workers.

@miq-bot add-label enhancement

[bdunne]

@jrafanie Would it make sense to tie these worker deployments to something like the CR rather than the orchestrator deployment? Would a new orchestrator be able to pickup and update the existing worker deployments?

[Fryguy]

I like the idea of being able to bounce the orchestrator independently from the workers. I think tying the workers to the CR would work.