Clean up the way logcontexts and threads work in the pushers

Clean up room alias creation

... and rename it, for even more sanity

Handle HttpResponseException more safely for federated groups

Wrap calls to deferToThread() in a thing which uses a child logcontext to
attribute CPU usage to the right request.

While we're in the area, remove the logcontext_tracer stuff, which is never
used, and afaik doesn't work.

Fixes #4064

Add information about the matrix-docker-ansible-deploy playbook

We don't do anything with the result, so this is needed to give this code a
logcontext.

This brings it into line with on_new_notifications and on_new_receipts. It
requires a little bit of hoop-jumping in EmailPusher to load the throttle
params before the first loop.

`on_new_notifications` and `on_new_receipts` in `HttpPusher` and `EmailPusher`
now always return synchronously, so we can remove the `defer.gatherResults` on
their results, and the `run_as_background_process` wrappers can be removed too
because the PusherPool methods will now complete quickly enough.

Each pusher has its own loop which runs for as long as it has work to do. This
should run in its own background thread with its own logcontext, as other
similar loops elsewhere in the system do - which means that CPU usage is
consistently attributed to that loop, rather than to whatever request happened
to start the loop.

simplifies the interface to _start_pushers

... and use it from start_pusher_by_id. This mostly simplifies
start_pusher_by_id.

This is public (or at least, called from outside the class), so ought to have a
better name.

Run MAU queries as background processes

Make psutil an explicit dependency

Fixes #3820

As of #4027, we require psutil to be installed, so it should be in our
dependency list. We can also remove some of the conditional import code
introduced by #992.

Fixes #4062.

This fallback didn't work, and was removed in #4069.

Calculate absolute path for email templates

Make manhole work on Python 3 again

Use the right python when starting workers

We should use the same python to start the workers as we do for the main
synapse (ie, the same one used to run synctl.)

this change has been released