4 months ago
We just faced this issue (as documented on https://station.railway.com/questions/networking-outbound-connection-problem-caafa330)
Angelo had told us to enabled static ip address and that this was due to intermeittment outbound-connectivity drop, and the pattern is always the same normal traffic flow -> 499 -> 502
Hi, is happening live to us right now, we have re-deployed 3 times, and it doesn't seem to be able to recover anymore.
23 Replies
This is the related ticket (https://station.railway.com/questions/networking-outbound-connection-problem-b3be7558)
4 months ago
hey there, all good, why I am taggable
4 months ago
working on this
4 months ago
so this is because of OOMing, not because of networking
4 months ago
cc @Lfret - are you around, it seems when it picks up a job from the queue its OOMing
4 months ago
are you using node for your backend?
yes we are but looking at the timestamp (Jun 30, 2026, 12:50:58 PM GMT+8 Hash: 97503f6e), there were 499 errors way before there was the OOM crash.
4 months ago
thats also related to the OOMing is my guess
4 months ago
uh, you are logging bitstreams?
4 months ago
redeployed with an expanded memory heap flag for you if thats okay, monitoring
4 months ago
bump
I think the bitstream is getting logged because our audio transcript job is failing, removing it now
4 months ago
okay, no crash
4 months ago
<:Fingers:1202140286956867644>
4 months ago
can you slam the backend for me?
4 months ago
which tz are you in?
4 months ago
are, we are out of the crash loop yea?
yes we are thank you so much angelo we will go and remove the OOM error now
4 months ago
np, least we can do