Beau Gunderson

I should note I'm using the snap-specific version of 2.0.1 with the default theme and the default font (Andale mono).

Looking for more logs because I know this isn't helpful as-is; sorry!

Was working fine all morning, then just starting spewing tons of "took too long to disconnect" errors; more stats: 

Also seeing a ton of "Exception inside application: Lock is not acquired." at the same time

Last time this state happened it was something in the redis db itself, as a `flushdb` fixed it--working on getting a dump of the db and will then `flushdb` and...

`flushdb` did indeed fix the issue; i'm looking through the dump now for clues

OK one weird thing is that my `zset`s for groups are huge, meaning they are not ever being expired... e.g. "asgi::group:account." contains ~100 entries, and for an organization, which all...

I wrote a script to generate these stats: ```python #!/usr/bin/env python import json import jsonlines def main(): # production.json is generated with redis-dump (`gem install redis-dump`) with jsonlines.open('./production.json') as reader:...

This is my `websocket_disconnect` handler: ```python def websocket_disconnect(self, message): logger.info(f'{self.account_key}: websocket_disconnect called\nmessage: {message}') if hasattr(self, 'presence_groups'): for group in self.presence_groups: PresenceGroup.objects.remove(group, self.channel_name) raise StopConsumer() ``` I'm using a custom implementation...

(Also noting that after the `flushdb` yesterday I hit this again this morning and it was fixed by another `flushdb`)