Workspaces are constantly disconnecting from the Server
Hi everyone 🙂
We're facing issues with our Coder deployment, hosted in a corporate network. We've been getting a lot of reports from users facing disconnections. We've tried some things - the most useful one was disabling STUN servers - but no solution seems to stick.
I'm attaching today's latency issues from Grafana, a screenshot from the health DERP page and the health endpoint response.
Can you give me some pointers on where should I improve our deployment or how to properly debug this issue?
3 Replies
we've had a similar issue recently but I don't know whether it's related with yours: https://discord.com/channels/747933592273027093/1307071966766694461
@vlvrd which version of Coder are you running?
also, is there a time where this issue occurs the most? and how often does it happen?
Coder v2.15.4
It happens very often, like every two or three hours now
and I can't match the behavior to a pattern except that it happens more often starting the work day and after lunch
Update: I used a PodAffinity rule to place the Coder server in the same node as postgres and that stabilized the network
oh okay
did that fix it completely ?