[Improvement]: split up docker-compose.yaml

[pimcore]

[Improvement]: split up docker-compose.yaml

[yariksheptykin]

Thanks for bringing up this discussion @passioneight . I also believe that docker-compose could benefit from some redesign.

I suggest we drop php-fpm-debug completely and instead parametrize php-fpm service to use either -debug or no-debug image flavor. This would also eliminate the need to "balance" between php-fpm-debug and php-fpm for nginx, which would simplify the config a bit. Triggering xdebug should better be implemented with xdebug means / triggers.

To adjust what ports DB-container binds to per environment it is indeed better to use docker-compose.override.yaml. We can create an example .override.yaml.dist in this repo to demonstrate some common customization. However, as this file is supposed to contain individual customizations I don't think we should add this file to the skeleton.

I think that trying to make this docker-compose.yaml staging- or production-ready is a good idea overall. It is, however, dangerous at the same time. This docker-compose.yaml should not be used as is in staging or production environments. Staging should align as close as possible with production. In most of the cases you would not use single-node compose based setups in production. I see the best fit for this docker-compose.yaml file is to set up production-near environment locally.

[jdreesen]

I suggest we drop php-fpm-debug completely and instead parametrize php-fpm service to use either -debug or no-debug image flavor. This would also eliminate the need to "balance" between php-fpm-debug and php-fpm for nginx, which would simplify the config a bit. Triggering xdebug should better be implemented with xdebug means / triggers.

I just created a PR which moves in this direction: #99