Watchdog stops SIPP from running

[tomer-amir-vonage]

Hi,

We built a SIP load testing tool for our servers, and we use SIPP for making calls. We run SIPP on AWS instances (c4.xlarge) running on Ubuntu, sometimes for long periods (hours). We also use PJSU to receive the calls made by SIPP, and PJSU is deployed on the same machine.

The loads we try to run do not exceed 150 concurrent sessions, at a rate which do not exceed 15 calls per second.

There are also some monitoring processes on the machines, but nothing which is CPU heavy.

The problem is that occasionally SIPP stops running because of the watchdog (both because of the major and minor thresholds):

22:16:08.030716	1488492968.030716: Overload error: the watchdog timer has tripped the major threshold of 3000ms too many times (201 out of 200 allowed) (866 out of 2400 minor 500ms timeouts tripped)

We tried to increase the watchdog thresholds, but it only delays the problem.

Is there something we can do to prevent it from happening?

Thanks!

[prabusampath007]

Facing the same issue, can someone help on this?