Use `IsBlacklistedOrBackingOff` from the federation API to check if we
should fetch devices.
To reduce back pressure, we now only queue retrying servers if there's
space in the channel.
This makes the following changes:
- Adds two new metrics observing the usage of the `DeviceListUpdater`
workers
- Makes the number of workers configurable
- Adds a 30s timeout for DB requests when receiving a device list update
over federation
This PR adds a config key `room_server.default_config_key` to set the
default room version for the room server.
Signed-off-by: `Sam Wedgwood <sam@wedgwood.dev>`
Also slightly refactors the functions and methods to rely less on the
req/res pattern we had for polylith.
Returns `M_WRONG_ROOM_KEYS_VERSION` for some endpoints as per the spec
Threepid seems to be pretty out of date, several missing endpoints.
Should also fix#3037, where we were still listening on the `/unstable`
prefix, while Element Web uses `/r0`
Adds some more checks in regards to power levels, uses a less heavy way
to get the membership of a user, avoids asking the database for the room
version, since it will be queried later. [skip ci]
This adds tests for `/profile`.
Also, as a first change in this regard, refactors the methods defined on
the `UserInternalAPI` to not use structs as the request/response
parameters.
