Please implement back off for matrix sync retries

[gaykitty]

Description

I run a Matrix server. We had a couple days of down time last week, during which we were getting absolutely spammed by sync requests from a single user. Over 2 million in 24 hours. After some sleuthing I figured out it was an instance of this bot sending the requests.

Steps to Reproduce

Run the bot against a matrix server that returns error when sync is called.

Expected Functionality

Implement some kind of back off if sync fails to multiple times in a row.

Experienced Functionality

Because our Matrix server was down, but our Nginx server that sits in front of it was still up, all requests immediately returned HTTP 502. So, every call to sync the bot made immediately returned an error, and the bot immediately tried the request again.

opsdroid/opsdroid/connector/matrix/connector.py

Lines 328 to 338 in e2ffb0c

	while True: # pylint: disable=R1702
	response = await self.connection.sync(
	timeout=int(60 * 1e3), # 1m in ms
	sync_filter=self.filter_id,
	since=self.connection.sync_token,
	)
	if isinstance(response, nio.SyncError):
	_LOGGER.error(
	f"Error during sync: {response.message} (status code {response.status_code})"
	)
	continue

This loop should implement some kind of back off when multiple error occur in a row to prevent this.

[RaitoBezarius]

I'm so sorry for that, I will implement exp back off ASAP and in the bot I run on your instance.

[gaykitty]

Thank you very much 💖

[Cadair]

I would be very happy to review a PR for this, but wont have time to write one myself. Sorry this project has got a little quiet recently, things are very busy for all of us.