18 Replies
we can't investigate with just that as info
but noted, if we have similar feedbacks we will use that as a note
let me know how I can help, provide a reproduction or some more detailed logs from app maybe?
Might have been due to still some old app pods hanging out, will let you know if it keeps happening.
It stopped 😛
Yes, most likely you're hitting a not authorized at some point which will log you out
if you can isolate what not authorized it is, it will help us
hmm nvm it keeps happening
which you can get from your server logs
If you can isolate how you got to make that request that will help us
ideally we can repro in a fresh isolated environment
I have a perpetual script that has a token attached to it, that is from my user.
That script checks if a flow it triggered has been completed.
So I guess that fails.
only requests you make from the browser will log you out
also this is status 400
you need to look for status 401
or status 403
{"timestamp":"2024-05-27T09:53:00.102601Z","level":"INFO","message":"response","latency":"3","status":200,"target":"windmill_api::tracing_init","span":{"method":"GET","uri":"/api/w/voja/jobs_u/completed/get_result_maybe/018fb974-9439-5aa4-3011-6c287f39e492","name":"request"}}
{"timestamp":"2024-05-27T09:53:00.118615Z","level":"INFO","message":"response","latency":"0","status":200,"target":"windmill_api::tracing_init","span":{"method":"GET","uri":"/_app/immutable/nodes/4.BDKe7_wv.js","name":"request"}}
{"timestamp":"2024-05-27T09:53:00.184896Z","level":"INFO","message":"response","latency":"0","status":401,"target":"windmill_api::tracing_init","span":{"method":"GET","uri":"/api/workspaces/users","name":"request"}}
I am not doing anything really, just log in, then get logged out intermittently
How often do you get disconnected ?
Every 30s or so
oh, then probably you have a misconfiguration somewhere in your ingress
cause otherwise windmill would be unusable for everyone
Didn't change configuration, went from 2.0.178 to 2.0.195 everything seems okay except this
Not seeing error logs around the 401, just these random infos
Yes but we have hundreds of enterprise customers with many many users themselves, and app.windmill.dev, that are on latest or close so. If it was our helm the issue, we would know
like every 30s is very very noticeable for everyone
so there is probably something wrong with your setup
if you can come back with more info we can investigate
fair point
This is the only error I see really
those are logs from the workers
alternatively, you can rollback to 2.0.178
and see if you do not have the error anymore
can try on a fresh instance as well, going to 2.0.178, testing, then 2.0.195, testing
Yep, thanks for the support 🙂