RTPAUDIOQOS all zeros

Guys, I have such problem.

I have test-environment:

  1. server with asterisk, that emulates VOIP-provider
  2. proxy-server (emulates case when provider cannot reach our main voip server directly)
  3. main voip server

I do test-calls for 2 scenarios:

  1. client call to us:
    zoiper (my PC) -> server #1 -> server #2 -> server #3 -> server #2 -> call-center agent (cisco hardphone)
    here server #2 also used as proxy server for call-center agents.
  2. call-center agent calls client:
    call-center agent -> server #2 -> server #3 -> server #2 -> server #1 ->zoiper (my PC)

I hear voice on both sides in both test-cases.
My problem is related to variables {RTPAUDIOQOS} and {RTPAUDIOQOSBRIDGED}

They work normally on server #1 and server #3 (I have asterisk 16 there).
But on server #2 they return values with all zeros, except “ssrc”, for example:

RTPAUDIOQOS ssrc=1816278608;themssrc=0;lp=0;rxjitter=0.000000;rxcount=0;txjitter=0.000000;txcount=0;rlp=0;rtt=0.000000
RTPAUDIOQOSBRIDGED ssrc=2070876554;themssrc=0;lp=0;rxjitter=0.000000;rxcount=0;txjitter=0.000000;txcount=0;rlp=0;rtt=0.000000

server #2 has asterisk version:
Asterisk 13.1.0~dfsg-1.1ubuntu4.1

We do not use PJSIP, just SIP, so it is chan_sip.c module.
In asterisk console “sip show channelstats” also gives all zeros during the call, example:
Peer Call ID Duration Recv: Pack Lost ( %) Jitter Send: Pack Lost ( %) Jitter
172.27.99.179 71cfe7a2300 00:00:08 0000000000 0000000000 ( 0.00%) 0.0000 0000000000 0000000000 ( 0.00%) 0.0000
172.27.99.224 c0f0f672-4d (inv state: None) – -- No RTP active
172.27.88.228 529d7a0d249 00:00:08 0000000000 0000000000 ( 0.00%) 0.0000 0000000000 0000000000 ( 0.00%) 0.0000
172.27.99.179 6ff9c7b0-24 (inv state: None) – -- No RTP active
2 active SIP channels

Can someone help me to understand why it might be; how to fix that?

Asterisk 13.1.0 is over 6 years old, so it is entirely possible there were issues in that version which were later resolved.