A coworkers shameless hack. Our services have a bug being investigated that cause the server eventually to fail.

The hack? Reset the connections of every user every 6 hours and restart the servers.

-_-

Wouldn't really call that shameless. That's fairly standard for those types of things. The devs need to prioritize fixing it but because the hack resolves it (albeit adding unnecessary burden to operations) they can't be bothered.

Still think this is a tech problem?

The system is a video streaming system, it causes all users about 5 minutes of downtime every 6 hours. Perhaps that isn't a huge amount, but there is quite often 5 minutes of news at midday for example :D.

Edit: it has also been like this for at least a month

Sounds like my team though we do it manually when we start getting angry emails from users saying it's slow. No one has tried to figure out the root cause though...