Priority based start and stop of heavy containers, specifically SOGo

[pquan]

Summary

I have an installation where the SOGo container will take more than 20 minutes to stop before updates. The rest of the containers are much faster, except for the start of the unbound container which takes around a minute.

Every time I update the mailcow installation, I have 20+ minutes of down town, where all containers are stopped and the SOGo container is "stopping" and waiting... and waiting... and waiting...

I'm suggesting that there's an option to stop SOGo (or any other container listed somehow/somewhere) before stopping the rest. At the same time, these containers shall be started at the end when restarting. The reason is to minimize the down time for most of the mailcow services, and specially for te SMTP/IMAP clients.

Motivation

As it stands, the down time for the mailcow update is about 20 to 40 minutes. This will allow the down time to be less than 1 minute, for most services, except web clients on SOGo.

Additional context

No response