If you’re running HA in a docker, you need to run additional containers for add-ons. This is called out in the docs. Add-ons are only for HA OS or if you install it natively, with the supervisor (HA Supervised).
If you are willing to dedicate a device to just HA you don’t need separate containers for the add-ons. For ease of use that makes a lot of sense, it’s, pretty plug and play.
Personally the Pi I’m running it on can handle a lot more than just HA so a docker makes more sense, and just have the add-ons I’m using also defined in the docker compose file.