Asterisk 16 - RINGNOANSWER with agent available after AgentRequest (Nobody picked up)

I would appreciate help with this problem:

In some cases (not all), agents are getting calls stuck in their internal queues (each agent has their own queue).

The agent is available, and the call should be automatically answered after the AgentRequest, but for some reason, Asterisk shows that it tried to call the agent and no one answered… I’m having this problem with both softphones and WebRTC.

This issue is generating some RINGNOANSWER records in queue_log, and prevents the agent from receiving the call.

Could someone help me with this? with the resolution or tips on how I could better investigate and debug this issue.

Here is some info I gathered:

queueagent123 has 1 calls (max 2) in ‘ringall’ strategy (14s holdtime, 45s talktime), W:200, C:48, A:0, SL:100.0%, SL2:100.0% within 120s
Members:
Local/agentcode123@join_queue/n (ringinuse disabled) (realtime) (Not in use) has taken 48 calls (last was 151 secs ago) (login was 7733 secs ago)
Callers:

  1. Local/queueagent123@internal-00002003;2 (wait: 0:12, prio: 0)

-VERBOSE[17454][C-0000e69f] pbx.c: Executing [agentcode123@join_queue:2] AgentRequest(“Local/agentcode123@join_queue-00002005;2”, “agentcode123”) in new stack
-VERBOSE[17454][C-0000e69f] bridge_channel.c: Channel Local/agentcode123@join_queue-00002005;2 joined ‘simple_bridge’ basic-bridge <9761c74d-0fbe-4a5f-ac85-0d6a1f5941dc>
-VERBOSE[17413][C-0000e69f] app_queue.c: Local/agentcode123@join_queue-00002005;1 is ringing
-VERBOSE[17413][C-0000e69f] app_queue.c: Nobody picked up in 7000 ms

→ extensions.conf

[join_queue]
exten => _.,1,Noop(callerid :: ${CALLERID(num)})
exten => _.,n,AgentRequest(${EXTEN})
exten => _.,n,Hangup()

Note: this issue is happening only where there is a high load of calls, but machine resources looks ok.

Update, I just found this error while debugging chan_sip:

chan_sip.c: FRACK!, Failed assertion bad magic number 0x7fe for object 0x7fe52827fba0