Inbound Trunk returning 408 Request Timeout - Livekit Cloud not responding to INIVTE

Telephony
1

Hi all,

We have a BYO carrier (Danish provider, plain UDP SIP trunk, no TLS) sending INVITEs to our LiveKit Cloud SIP endpoint. Symptom: zero SIP response at all — not even 100 Trying — across multiple retries to two different resolved edge IPs (both within LiveKit’s documented static SIP range). After ~5s the carrier’s own proxy gives up and returns a 408 on its side. That 408 is self-generated by the carrier, not something LiveKit sent.

We’ve tried various combinations of inbound trunks and dispatch

  • filtering on their phone no’s
  • Without phone number filtering, allowing using 0.0.0.0

From the Provider-wireshark, it looks like the Invite is never picked up despite hitting the livekit cloud SIP IP.

Happy to share project/trunk IDs with a staff member directly if that helps track this down - didn’t want to post them in the open here.

Wireshark from provider:

Group 11 (2)
Group 11 (2)1260×374 62.9 KB

Any ideas on how to best debug this?

2

Do oyu have a call id or PCAP or something I can look at?

3

I found an old Slack thread that pointed to differences in the SIP URI using IP vs FQDN, and asked provider to adjust accordingly and that seem to have done the trick :slight_smile: