We were only reporting failures for when the async task didn't finish.
We also need to report a failure for when the task finished, but
returned a non-zero exit code.

This is more consistent with how we name variables. It's also less
confusing, especially given that we have `matrix_hookshot_feeds_pollTimeoutSeconds` as well.

ref: https://github.com/mautrix/signal/releases/tag/v0.4.1

Fixes https://github.com/spantaleev/matrix-docker-ansible-deploy/issues/2211

Signed-off-by: Sebastian Gumprich <github@gumpri.ch>

ref: https://gitlab.com/signald/signald/-/issues/343

`matrix-postgres-backup` reuses the `matrix-postgres` role for these
tasks, not its own definitions.

Fixes https://github.com/spantaleev/matrix-docker-ansible-deploy/issues/2200

At some point, we can drop v9.6 support and use a single variable
that holds this value.

Fixes https://github.com/spantaleev/matrix-docker-ansible-deploy/issues/2191

fixes jitsi issue

ref: 57bb340343

We're starting it with `--user` and dropped capabilities, after all.

Hopefully fixes https://github.com/spantaleev/matrix-docker-ansible-deploy/issues/2188

We haven't encountered such a problem yet, but it doesn't hurt to
make things more robust.

These values that we were setting also make sense in the context of the
`matrix-postgres` role even when not used within the playbook.

Related to https://github.com/spantaleev/matrix-docker-ansible-deploy/pull/2180

Just specifying `15` means we won't automatically re-pull `15.1` when it comes out.

We're overriding these in the correct way in `group_vars/matrix_servers`
so this wasn't causing any problem in practice.

.. when `matrix_nginx_proxy_enabled: false`

Fixes https://github.com/spantaleev/matrix-docker-ansible-deploy/issues/2177