alternative to supervisord in 2.11.x for ckanext-harvest

After supervisord has been removed from 2.11.x images without replacement, what is the recommended approach to integrate ckanext-harvest into the containers again?
Currently it breaks existing functionality when updating from 2.10.x containers.

Should one trigger gather and fetch jobs once via docker healthcheck?
Running it in background using nohup?
Trigger them from outside via host cron?

Or should i go for implementing supervisord manually again?

Would like to collect some ideas, thank you. :-)

No files, tests, or entry points are named. Start by reviewing the 2.11.x image configuration and how ckanext-harvest gather and fetch jobs were previously started; compare the healthcheck, nohup, host cron, and supervisord options. Done requires an agreed replacement approach that restores scheduled harvest jobs in the containers.

Written by the indexing model from the issue text.