Hi,
We’re getting 404 No trunk found on every inbound SIP call from Zadarma (Bulgarian SIP provider) to our LiveKit Cloud project. After extensive debugging we believe this is a GeoDNS/anycast routing issue similar to the one described here: SIP 404 “No trunk found” — DNS routing mismatch on LiveKit Cloud
Environment:
-
LiveKit Cloud project SIP URI:
ytlj4qounpm.sip.livekit.cloud -
SIP provider: Zadarma (zadarma.com) — Bulgarian virtual number
-
Inbound trunk ID:
ST_o3tpcif66zzu -
Dispatch rule configured with
agentNamematching a running Python agent
What the SIP log from Zadarma shows:
To: <sip:35924373387@ytlj4qounpm.sip.livekit.cloud>
SIP/2.0 404 No trunk found
Via: SIP/2.0/UDP 185.45.152.192:5060
src: 185.45.152.200 (Zadarma PBX)
dst: 161.115.161.184:5060
What we confirmed:
-
TCP port 5060 open and responding

-
nslookup ytlj4qounpm.sip.livekit.cloud→161.115.161.26and161.115.161.184
-
Zadarma IP ranges correctly added to
allowedAddresses
-
LiveKit → Telephony → Calls shows 0 calls — never registers anything
-
Call forwarding to a real phone number works perfectly

-
Python agent running and registered as
bulgarian-agent
Trunk configurations tested (all return 404):
-
With
numbers: ["+35924373387", "35924373387"] -
With
numbers: [](empty — accept any number) -
With and without
allowedAddresses -
With and without
authUsername/authPassword
Suspected cause: Zadarma sends from Bulgaria and likely hits a different LiveKit SIP server than the one where our trunk is registered — same pattern as the DNS routing mismatch thread above.
Question:
-
Is
ytlj4qounpm.eu.sip.livekit.clouda valid EU region-pinned endpoint we can use? -
Will region pinning solve the GeoDNS mismatch?
-
Is there anything else in the trunk configuration we might be missing?
Thanks!