This is an attempt at optimizing service startup.
The effect is most pronounced when many services are restarted one by one.
The systemd service manager role sometimes does this - for example when `just install-service synapse` runs.
In such cases, a 5-second delay for each Synapse worker service
(or other bridge/bot service that waits on the homeserver) quickly adds up to a lot.
When services are all stopped fully and then started, the effect is not so pronounced, because
`matrix-synapse.service` starts first and pulls all worker services (defined as `Wants=` for it).
Later on, when the systemd service manager role "starts" these worker services, they're started already.
Even if they had a 5-second wait each, it would have happened in parallel.
- forego removing Docker images - it's not effective anyway, because it
only removes the last version.. which is a drop in the bucket, usually
- do not reload systemd - it's none of our business. `--tags=start`,
etc., handle this
- combine all uninstall tasks under a single block, which only runs if
we detect traces (a leftover systemd .service file) of the component.
If no such .service is detected, we skip them all. This may lead to
incorect cleanup in rare cases, but is good enough for the most part.
These `init.yml` (now `inject_into_nginx_proxy.yml`) tasks do not need
to `always` run. They only need to run for `setup-all` and
`setup-nginx-proxy`. Unless we're dealing with these 2 tags, we can
spare ourselves a lot of work.
This patch also moves the `when` statement from `init.yml` into
`main.yml` in an effort to further optimize things by potentially
avoiding the extra file include.
