`rtp_timeout` log message

ssherar · March 21, 2024, 12:36pm

Hi

Looking to tweak our rtp_timeout and rtp_timeouthold values from being disabled to a more sensible number to handle potential zombie calls. However, with the volume of calls going through, we will more than likely end up only hearing about problems from customers as there is no easy way to reproduce any issues that customers might have

So instead, I would like to monitor the log messages to see which channels have been terminated due to lack of RTP traffic through our centralised logging systems.

Is there a message (or message pattern) that signifies that a channel has been terminated by an RTP timeout and if so, can someone provide a sample?

jcolp · March 21, 2024, 12:38pm

github.com

asterisk/asterisk/blob/master/res/res_pjsip_sdp_rtp.c#L146


      
          		*/
          		return timeout * 500;
          	}
          
          	elapsed = now - ast_rtp_instance_get_last_rx(rtp);
          	if (elapsed < timeout) {
          		ast_channel_unref(chan);
          		return (timeout - elapsed) * 1000;
          	}
          
          	ast_log(LOG_NOTICE, "Disconnecting channel '%s' for lack of %s RTP activity in %d seconds\n",
          		ast_channel_name(chan), ast_codec_media_type2str(session_media->type), elapsed);
          
          	ast_channel_lock(chan);
          	ast_channel_hangupcause_set(chan, AST_CAUSE_REQUESTED_CHAN_UNAVAIL);
          	ast_channel_unlock(chan);
          
          	ast_softhangup(chan, AST_SOFTHANGUP_DEV);
          	ast_channel_unref(chan);
          
          	return 0;

ssherar · March 21, 2024, 1:57pm

Perfect - thanks @jcolp!

system · April 20, 2024, 1:57pm

This topic was automatically closed 30 days after the last reply. New replies are no longer allowed.