Questions about the supervisor config

[danielballan]

Reviewing a0c9e2b#diff-8eedc134d422e0ea27511b9854a9911105811b0baecc17dd98554b01f095c3e7

1. Normally I would expect we'd just need a file with [program:...] block that should be placed in .../conf.d (where ... varied by platform / how supervisor was installed). But these files are named supervisord.conf and supervisorctl.conf and seem to configure supervisor itself as well as the bluesky-kafka consumer program. I'm not familiar with some of the directives and the commit message is not specific, so I can't tell what it's doing. Can you @jklynch or @gwbischof comment on what these are doing?
2. These aren't going to be used by the Python package, so the supervisor directory should be lifted into the root directory of the repo, outside of the Python package distribution.
3. Rendering (1) and (2) possibly moot, I've come around to thinking that we should listen to @stuartcampbell and just use systemd. It's only slightly more complicated than supervisor and it's a much more solid (and performant) system tool. Our use of supervisor is to some degree due to me, and probably an artifact of my "Python all the things!" blinders. :- ) We could keep both configurations here if we want, so people have options.

[jklynch]

My interpretation of the evidence is that we wanted to save working supervisor configurations that we were using during development, and that we were not skillful supervisor users (I still am not skillful). I don't think we need to keep these files anymore, but @gwbischof may feel differently. Also, the mongo_consumer code is slated to move to a different repository.

Personally I am happy to move from supervisor to systemd, but I think we can just remove the supervisor configuration files entirely from this repository without replacing them with equivalent systemd files.