Issue with RTMP ingest to Akamai

Hi,

We are transcoding RTP/RTSP input into 3 resolutions (720p, 360p, 240p) and using RTMPPushPublish pushing the stream to Akamai entrypoints. For some of the streams being published, Akamai is complaining that the RTMP stream being pushed to akamai entrypoint has out of order timestamps (PTS is out of order). We looked at the wowza logs and didn’t find anything related to timestamps or PTS of streams being out of order. When we connect to wowza output stream as an rtmp client, the stream plays fine. But the same stream being pushed to akamai over RTMP seems to be having timestamp issues as reported by akamai.

Here is the link to pcap file that captures network traffic between wowza and akamai: https://bluejeans.app.box.com/s/6vku6qfc5szdy69s3fkz786fibvh3cpf . Can you look at the pcap files and see if the timestamps in the RTMP stream being pushed to akamai are out of order?

Thank you,

Rakshith

Hello Rakshith,

I believe we have handled this issue in the ticket that is in our system, if there is more we can do to assist you in this public forum please let us know.

Regards,

Jason Hatchett

Hi, I am having the same exact issue with Akamai ingest. Could you please share the solution?

Any possibilities to get solution shared to us too.

We might have similar issue while ingest is not RTP/RTSP but MPEG-TS over UDP(multicast) as ingest to Wowza and then encoded to 3 flavours with wowza transcoder. Everything plays fine except when sent to Akamai RTMP ingest (MSL3.2 stream packaging to HLS). Issue is manifesting itself to Akamai around/before every 46hours stream being pushed to Akamai.

Problem is shown in very laggy video as lot of dropped frames and audio getting cutoff quite badly. While this situation is present in Akamai HLS packaged stream same ingest/transcode plays fine from that pushing Wowza (RTMP and HLS packaged) and even same streams pushed to separate Wowza instance are playing fine (Though that RTMP push were not pushed same amount of time to other wowza instance).