We have a number of Wowza Streaming Engine (4.8.5 / JRE 9.0.4) instances running on dedicated servers (Ubuntu 20.04) pushing transrated HLS streams to Wowza Cloud CDN Stream Targets. Each server handles a number of applications and on occasion (increasingly so in the last few weeks) applications will randomly not send streams to the CDN. If I check the application during the incoming RTMP ingest, the Incoming Stream is working and recording fine, the transcoder is working fine (I can check the renditions), but the Stream Target is stuck in the “Waiting” mode. There are only 2 fixes:
Restart the application - not ideal since we get a break in the recording
Delete and re-add the Stream Target - also not ideal, since we have to then log into Wowza Cloud, regenerate connection codes and go through the re-linking process
Once one of those two fixes is done the Stream Target immediately goes “Active” and proceeds to work fine. There is no pattern or predictability to this issue - it can happen on any of the 160+ applications we have across 30+ dedicated servers and the above two fixes are the only solution. I have checked the logs on the affected servers / applications when I catch this problem and there are no errors indicated at all.
Would really appreciate some help with this one.