-
Notifications
You must be signed in to change notification settings - Fork 79
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Request for Help: WebSocket Connection Issue with Jitsi on Azure Kubernetes #112
Comments
colibri websocket is a permanent connection between the client and the server. If ingress drops the connections after a while, this will break the signaling. You may try to increase timeouts:
|
Hi emrahcom, Thanks for the suggestion! I tried updating my configuration as you recommended, but unfortunately, it didn't resolve the issue. The error persists, and it seems like the WebSocket connection isn't being established at all. Any other ideas or adjustments I could try? Appreciate your help! |
Hello @Bananenbrot1! You can try using the current # jitsi-values.yaml
websockets:
colibri:
enabled: false
xmpp:
enabled: false Current If it works for you, you can keep using this version or revert to the latest release and add these options to your values: # jitsi-values.yaml
extraCommonEnvs:
ENABLE_SCTP: 'true'
ENABLE_COLIBRI_WEBSOCKET: 'false'
JVB_PREFER_SCTP: 'true' |
It may be also related with It should be the following for this case:
|
Hi @spijet, Thanks for the quick response! I implemented the extraCommonEnvs as you suggested, and it indeed resolved the error. Now, I can successfully receive and stream videos at 360p, which is a small but appreciated improvement 👍. However, I'm hesitant about using a legacy version since I'm aiming to build a clean, modern stack. Do you have any insights on how I might enable the websockets to function properly in this setup? Do you think upgrading to the head of main of the chart would solve the issue? Hi @emrahcom, Thank you as well for your prompt reply. I tried setting the environment variable as you advised, but unfortunately, there's been no change—I'm still seeing the same error message. Any other suggestions? Thanks again for all the help! |
It's "legacy" in the sense that it's been there before the WebSockets support was implemented. IIRC, meet.jit.si currently uses both WSS and REST/SCTP in order to maximize the support among different client browsers and apps. As for the 360p video resolution — make sure that you're not looking at the video feed from mobile clients, as they're known to limit their video upload to 360p max, in order to conserve traffic/bandwidth and go easier on the battery (see jitsi/jitsi-meet#5808 (comment) for details). Can you please share the full error text from the browser console and complete contents of your chart values, with sensitive info (like usernames/passwords/urls/ips) masked? |
Hey @spijet sure! Here is our values.yaml file.
And here is the browser console log:
|
Your values seem to be OK, apart from the As for that WSS error, code 1006 seems to mean that the local side closed the WSS connection for some reason. Just in case, can you check if you can reach the URL mentioned in the error with Another possible cause is that Azure Ingress controller needs some special configuration in order to start passing the WSS connections through. I haven't had any experience with AKS yet so I don't know if that is the case, but it can't hurt to check. |
Hey @spijet, I wanted to update you about the Octo configuration—we conducted a load test yesterday evening with 110 users across two video bridges, and it worked flawlessly! The HTTPS link is accessible, so no issues there. Regarding AKS, I initially thought that might be a problem area, but I haven’t found anything conclusive yet. I'll continue to investigate. Thanks so much for all your help! |
I conducted further testing, and it turns out the 360p issue was due to Jitsi automatically scaling down the video for conferences with more than three users, which makes sense. @spijet, I also tested with three video bridges and three participants. Each participant was connected to a different bridge, and the conference quality was very good. I then simulated a failure by shutting down two bridges, and all users were seamlessly reassigned to the remaining active bridge. Everything worked perfectly. I'm going to close this ticket now. Thanks for all the help! If I find a solution for the websocket issue, I'll update this thread. |
I haven't heard about it until today. I assume you've tested it in tile mode? If you switch to the default "one big talking head + bunch of smaller ones", the "big" video should be able to scale up to 720p, while the rest will be limited to 180p. Anyway, glad to hear that OCTO works flawlessly for you! :) |
Hi , pleas can you tell me how did you solve your problem with AKS and websocket? |
Hi everyone,
Firstly, I want to commend the excellent work on the Chart! It's been incredibly helpful.
I'm currently facing an issue with my Jitsi setup on an Azure Kubernetes Cluster where WebSocket support, intended for higher video quality, consistently fails. Here's the error I'm encountering:
To provide some context, the Jitsi Video Bridge (JVB) is publicly available via hostport. Additionally, here’s the ingress configuration for web access:
Could anyone suggest any modifications to the ingress settings or other configuration changes that might help resolve the WebSocket connectivity issue? I'm wondering if specific adjustments are necessary for the ingress to support WebSocket connections.
Thank you for your time and any assistance you can provide!
The text was updated successfully, but these errors were encountered: