helm: general cleanup

[derekperkins]

This is a lot of cleanup work that feels good to have done. TL;DR:

1. I changed all the imagePullPolicy references to be Always. This was set inconsistently and caused some issues with people trying it out. If this causes too many pulls in production, we can add a user flag for it later.
2. I switched out a few logging containers that I forgot to handle in helm: logging improvements #4343
3. The InitShardMaster job wasn't getting the proper labels
4. Setting init_db_name_override=keyspace.name is the most impactful change since I am changing a default. I did this because many database clients want to prefix the database name on calls. By making keyspace name == schema name, this allows those to work properly. I did add a setting to values.yaml to turn it off for compatibility.
5. There have been formatting issues where too much whitespace was being removed, so I reduced the amount of whitespace reduction that go templates are handling.

[dkhenry]

For setting the imagePullPolicy I think we should remove it everywhere and allow the users to use the default behavior of pulling when the tag is set to latest this would allow a user to specify a tag for the images and give them the option of using discreet images

[derekperkins]

There's already support for setting custom image tags, just not the imagePullPolicy. I still think the default for the helm chart should be Always, simply because there have been multiple issues reported that would have been solved by that. With the exception of maybe vtgates in a spiky workload with an HPA, these containers aren't really repulling images all that often anyways.

I agree that imagePullPolicy should be settable, my vote is to do that in a future PR.

[dkhenry]

LGTM