examine large chown'ed directories in container images and see if they can be reduced

[mmguero]

sparked from #861 and things like containers/podman#16541

In the entrypoint script shared by all containers there are some chown operations that have to occur to make sure files/directories are owned by the right UID/GID. When the user running Malcolm (PUID/PGID in process.env) matches what was there at build time (1000/1000 by default) this isn't too expensive. But when the user is different, it can take a long time, particularly on rotational backing media.

Some of this can't be avoided. However, I think that there are some large directories (large as in numbers of files, not file size) that are being blanket-chowned that we could probably be more targeted on.

zeek-live, logstash, dashboards, and netbox seem to be some of the worst offenders.

We ought to:

1. for testing purposes (not committed to GitHub) add timing output to the running of docker-uid-gid-setup.sh
2. determine the ones that take the longest (as a non-1000 uid/gid user)
3. figure out which directories/subdirectories could be excluded from this, if possible

[divinehawk]

Noticed this: zeek:26.02.0 is uid 0 in /usr/local/zeek, so even 1000 gets a costly chown... at least 15 minutes on my system (not SSD).

[mmguero]

I recently changed base images to the docker image, and it could be that I need tweak that. thanks for that feedback.