need to improve timeout control on sync-openstack-images

A few things aren't right:

• I noticed a run where sync-openstack-images Job took 35 minutes to run -- this is just too short compared to the HelmRelease timeout
• the HelmRelease.timeout isn't consistent with values.job.activeDeadlineSeconds
• activeDeadlineSeconds is computed based on the full list of images (which is very long), while ideally it should be based on the number of activated images (images matching a selector from os_image_selectors)

The ideal solution isn't easy to find:

• we can compute activeDeadlineSeconds in the sync-openstack-images chart to let it compute an estimate based on the number of activated images (images matching a selector from os_image_selectors)
• we can't align the HelmRelease.spec.timeout on this, because when the HelmRelease is generated, we don't have the results of this computation
• what we can do, perhaps is compute HelmRelease.spec.timeout based on the number of selectors ...