Deleting stale devices causes lots of device updates to be sent out over federation

[anoadragon453]

#14595 introduced a background job code that runs in DeviceWorkerHandler.check_device_registered to remove old, non-e2ee devices for users that had >50 devices, in an attempt to reduce the number of devices on homeservers.

When this was applied to matrix.org, it ended up spamming homeservers over federation with device list updates.

We may want to consider a rate-limit for the amount of devices we remove in this - and future - cleanup jobs.

[anoadragon453]

Was it expected that this code should run when a user uploads e2ee keys?

synapse/synapse/handlers/e2e_keys.py

Lines 668 to 673 in 6d47b7e

	# the device should have been registered already, but it may have been
	# deleted due to a race with a DELETE request. Or we may be using an
	# old access_token without an associated device_id. Either way, we
	# need to double-check the device is registered to avoid ending up with
	# keys without a corresponding device.
	await self.device_handler.check_device_registered(user_id, device_id)

If not, it may be causing much more frequent device pruning than expected. We may want to prevent the stale device purge code from running in this case.