VoIP high availability and failover

High availability is only real when the failure mode is known, detection is fast enough, and the alternate path has been tested under realistic load.

Failure domains to consider

• Power, switching, Wi-Fi, and local cabling
• Internet circuits and upstream routing
• DNS, certificates, and identity services
• PBX, SBC, recording, and provisioning systems
• Carrier routes and number presentation
• RTP reachability and NAT state

Document who detects the failure, how calls in progress behave, where new calls route, and how operators know which path is active. Test recovery as well as failover.

Related: SIP trunking, DNS/SRV, and monitoring.