The behaviour was changed with #318 . DNS is typically a critical part of the infrastructure - in case of power outage, blocky will be restarted at some time later and should run even if database is down, otherwise you'll have no DNS. Maybe we should add some kind of retry: for example blocky tries 3 times to establish the db connection with 3 seconds pause between attempts?

If you use a single compose file with blocky and mariadb, you can try to add depends_on with "condition: service_healthy" for mariadb. In this case, docker will start mariadb first and if the service is healthy (health check), blocky will be started.

Right, makes perfect sense. I'll try to fix it with my docker setup, thx for the suggestion, but retries sounds like a good idea anyway since my setup probably isn't the only were stuff doesn't always work on the first try.

Also, thanks for making this cool project, I've sent you a coffee (or whatever you want) on PayPal.