This is a bit of a compatibility break.
The role was defaulting the Postgres password to `some-password` and we
auto-generate it now.

However, rebuilding both Postgres and this service should unify the
database credentials and the service configs to the new value.

Same as 250b91a40968e, but for Slack

Postgres is not in `matrix_docker_network` anymore, so what we had
before could not possibly work anymore.

This requires at least `v0.14.0-2` of the `prometheus-postgres-exporter`
Ansible role.

This requires at least `v1.7.0-2` of the `prometheus-node-exporter`
Ansible role.

This is what we do for Synapse as well.

Fixes https://github.com/spantaleev/matrix-docker-ansible-deploy/issues/3089

Such a core service probably belongs better when it's in the homeserver network

This is mostly so as to avoid referring to variables from other roles,
like `matrix_nginx_proxy_proxy_matrix_client_api_addr_with_container`.

This should come (and already does) from group_vars/matrix_servers

This also makes it handle the `/_matrix/client/v3/register` endpoint,
not just `/_matrix/client/r0/register`

This reverts commit bf95ad2235.

This was a bad idea.
It's better to have people manually define the password.

Otherwise, `matrix_homeserver_generic_secret_key` changing some day in
the future would break the bot and one would have to figure out how to
reset its password manually.

Using an explicit password is more stable.

This is split out from matrix-cactus-comments (see 241779b583),
but also heavily inspired by `matrix-static-files`.

The container image has only ever been available for amd64,
so not enabling self-building for the other architectures was a mistake
that orignally landed in:
https://github.com/spantaleev/matrix-docker-ansible-deploy/pull/2089

Related to https://github.com/mautrix/discord/issues/95