We have encountered a failure where the session ou...
# egress
r
We have encountered a failure where the session output unexpectedly stopped. Not sure what the issue was. - Project: p_4rj9mtyggok - Egress ID: EG_vPMRLCUCDytz - Timestamp of Failure: 18 Dec 2025, 173510
c
Hey @ripe-lunch-17907 based on logs I can see the recording stopped because Stop API has been invoked - I would suggest checking what triggered it on your side if it was unexpected
that happened 4mins later comparing to the timestamp you provided though (assuming your timestamp above is Indian time - IST) - did stop happen explicitly after the output was broken? There is nothing in logs to indicate an issue during these last few mintutes
r
@clever-nail-49292 thanks for the prompt response. So the session was ended using Stop Api but the video out was crashed before session end. Please check this video at 10:00 minutes for reference
👍 1

https://www.youtube.com/watch?v=XqXErvGlIMA▾

c
I looked at chrome logs - this was chrome renderer crash - and based on these logs it's possible that it was caused by memory pressure. I see repeated lines indicating that
MediaRecorder
is active and that produces tens of MBs of webm blobs in rapid succession. This alone could be the source of the issue - what's the reason for having
MediaRecorder
enabled for the egress URL?
r
@clever-nail-49292 can you please share all the logs related to this egress
c
shared some snippets in the DM
We can updating egress to try detecting how frequently these happen so that we are aware of the impact as the first step
r
Is these any signalling mechanisms in web egress like room composite to know if it is active or not
c
you mean webhooks?
r
No not webhooks, I m talking about the one room composite recording uses for starting the egress after any participant joins the room. Something like this so we can reload the chrome if it is not active
c
I would need to double check - but I am pretty sure web egress doesn't have something like that
r
@clever-nail-49292 Hey, what event will be triggered for the pipeline frozen issue? Is it egress_updated? How to know it is related to a frozen pipeline?