... since the whole response is huge.

We even need to break up the assertions, since kibana otherwise truncates them.

This makes it easier to use in an async/await world.

Also fixes a bug where cache descriptors would occaisonally return a raw
value rather than a deferred.

Track the time that a server started failing at, for general analysis purposes.

This gives a bit of a grace period where we can attempt to refetch a
remote `well-known`, while still using the cached result if that fails.

Hopefully this will make the well-known resolution a bit more torelant
of failures, rather than it immediately treating failures as "no result"
and caching that for an hour.

There was some inconsistent behaviour in the caching layer around how
exceptions were handled - particularly synchronously-thrown ones.

This seems to be most easily handled by pushing the creation of
ObservableDeferreds down from CacheDescriptor to the Cache.

- Put the default window_size back to 1000ms (broken by #5181)
- Make the `rc_federation` config actually do something
- fix an off-by-one error in the 'concurrent' limit
- Avoid creating an unused `_PerHostRatelimiter` object for every single
  incoming request

when processing incoming transactions, it can be hard to see what's going on,
because we process a bunch of stuff in parallel, and because we may end up
recursively working our way through a chain of three or four events.

This commit creates a way to use logcontexts to add the relevant event ids to
the log lines.

ExpiringCache required that `start()` be called before it would actually
start expiring entries. A number of places didn't do that.

This PR removes `start` from ExpiringCache, and automatically starts
backround reaping process on creation instead.

Because it was complicated and annoyed me. I suspect this will be more
efficient too.

Linearizer was effectively a Limiter with max_count=1, so rather than
maintaining two sets of code, let's combine them.

Fixes #3570

a61738b3 removed a call to run_on_reactor from a unit test, but that call was
doing something useful, in making the function in question asynchronous.

Reinstate the call and add a check that we are testing what we wanted to be
testing.

When _get_state_for_groups is given a wildcard filter, just do a complete
lookup. Hopefully this will give us the best of both worlds by not filling up
the ram if we only need one or two keys, but also making the cache still work
for the federation reader usecase.

So, it turns out that if you have a first `Deferred` `D1`, you can add a
callback which returns another `Deferred` `D2`, and `D2` must then complete
before any further callbacks on `D1` will execute (and later callbacks on `D1`
get the *result* of `D2` rather than `D2` itself).

So, `D1` might have `called=True` (as in, it has started running its
callbacks), but any new callbacks added to `D1` won't get run until `D2`
completes - so if you `yield D1` in an `inlineCallbacks` function, your `yield`
will 'block'.

In conclusion: some of our assumptions in `logcontext` were invalid. We need to
make sure that we don't optimise out the logcontext juggling when this
situation happens. Fortunately, it is easy to detect by checking `D1.paused`.

This was introduced in 4f2f5171

This doesn't feel like a wheel we need to reinvent.