Support gunicorn worker recycling

[markafarrell]

Currently the memory usage of the multi-process container increases over time the majority contributor is the gunicorn workers for pulpcore-api

The suggested fix for these type of issues seems to be to enable worker recycling.

pulp/pulpcore/issues/2250

The suggested solution of editing gunicorn.conf.py, however, does not work (I think due to how s6-overlay works)

This changes adds some additional environment variables that can be use to configure gunicorn worker recycling for both pulpcore-api and pulpcore-content

[mikedep333]

Awesome, thank you so much!

[mikedep333]

@markafarrell Lint is failing.

You need a file CHANGES/576.feature on your commit that says "fixes #576"

[markafarrell]

@markafarrell Lint is failing.

You need a file CHANGES/576.feature on your commit that says "fixes #576"

Fixed.... hopefully

[markafarrell]

CI is failing due to:

 The conflict is caused by:
    The user requested pulpcore 3.45.0.dev0 (from git+https://github.com/pulp/pulpcore.git#egg=pulpcore)
    pulp-ansible 0.22.0.dev0 depends on pulpcore<3.55 and >=3.40.0
    pulp-certguard 1.8.0.dev0 depends on pulpcore<3.45 and >=3.28.0

pulp-certguard's pulpcore version restrictions need to be expanded

pulp/pulp-certguard#352
pulp/pulp-certguard#353

[markafarrell]

 pulp-oci-images CI / app-images (stable, pulp, pulp-web, pulpcore) (pull_request) Failing after 19m

This also appears to be failing on latest

#579

[mikedep333]

@markafarrell Yeah, we'll fix both of those 2 issues soon.

Thank you!

[mdellweg]

This change is incompatible with (at least) pulpcore 3.34. I'm inclined to revert it.

[mdellweg]

Nightlies are failing for some days now. We need to revert this and find a better solution.
https://github.com/pulp/pulp_gem/actions/runs/7620480802/job/20755309808#step:10:558