There are some weird issues, with hassio installed on debian buster.
Sometimes when I restart the homeassistant container for an update/configuration change(docker restart homeassistant)
The container cannot resolve dns anymore(neither internal hassio or exterior).
I tried restarting the supervisor service. checked logs etcetc
The only way to fix this is rebooting the entire server. I cannot find the underlying issue, maybe someone can point me in the right direction?
You need to restart the supervisor container and the Home Assistant container… perhaps just the supervisor.
I also restart the supervisor service. Most of the community addons have to be manually restarted from the hass.io menu.
This happens every time there is a docker update unfortunately.
Worst case a reboot required. (I have exactly the same setup)
Other containers using docker-compose and restart-unless-stopped all restart themselves it hass.io…
Tried restarting the supervisor before. Only option was rebooting.
It seems that something is breaking in the hassio networking setup.
The real issue is including hassio in an automated backup system, with a local sqllite db or mariadb addon.
Since that requires to stop - backup - start homeassistant/mariadb container, to safely backup the db without corruption.
1/2 Times hassio can’t find its dns anymore, shortly after aggro towards me occurs b/c ‘The Thing broke AGAIN!’ 
Guess best option will be to move the database out of hassio environment, or automate a mysqldump from the db.
I’m using mariadb on the same computer (nuc) as Hass.io without any issues. I use Portainer normally ro restart home assistant and supervisor containers. I’ve never had any problems doing that except some addons have to be restarted through the Hass.io side panel as using Portainer they just stop again straight away - mostly the community addons seem to do this.