Description of problem:
When osbuild-composer package is updated on a RHEL system where osbuild-composer is already running, the upgrade process results in the local worker to be killed and not started again. The result is that while osbuild-composer is running and responds to API requests, none of the jobs are ever picked by a worker, because there is none running.

Version-Release number of selected component (if applicable):
osbuild-composer-62-3.el8_7

How reproducible:
always

Steps to Reproduce:
1. dnf install y osbuild-composer composer-cli <- e.g. on RHEL-8.6
2. systemctl enable --now osbuild-composer.socket
3. composer-cli status show
4. systemctl status osbuild-worker@1.service
5. enable the latest RHEL-8 repos
6. dnf upgrade y osbuild-composer <- scriptlets take long time and it seems that systemd times out and kills processes
7. composer-cli status show
8. systemctl status osbuild-composer.service
9. systemctl status osbuild-worker@1.service <-- inactive (code=killed, signal=TERM)

Actual results:
osbuild-worker@1.service is inactive after RPM package upgrade

Expected results:
osbuild-worker@1.service should be running after RPM package upgrade, if it was running before the upgrade

Additional info:
Workaround is to start the worker service manually using:
systemctl start osbuild-worker@1.service