The intermittent issue with agents not answering Twilio Inbound calls—where most calls ring out but some connect—is a known scenario and is often related to configuration and infrastructure reliability. • Primary recommendation: For high reliability, you must run multiple servers for each agent name. This redundancy significantly reduces the chance of calls ringing out without assignment. • Resource isolation: Ensure your agent servers are dedicated—do not run other applications on these servers, as resource contention frequently causes intermittent failures. • Timeout mechanism: The system has a 10-second room connection timeout. If your agent does not establish a connection within this window, calls will ring out and not be assigned. • Follow production guidelines: Review and implement best practices such as proper load balancing, auto-scaling, and ensuring constant worker availability. This aligns with production deployment standards and mitigates these issues. • LiveKit built-in failover: LiveKit will attempt to assign the call to another available worker if one fails, but without sufficient reliability and failover configuration, unassigned calls can still occur. Adjust your deployment to include redundant agents, isolate resources, and follow infrastructure guidelines to minimise these intermittent failures. If you have already implemented these steps and still face the problem, or if you need guidance on specific configuration details, please provide more context about your deployment setup or any relevant logs/errors. Full reference details available upon request. Sources: How to Get Help from LiveKit | Speech & Audio Processing | livekit/livekit_composite | DeepWiki | pkg/sip/inbound.go | Web and mobile frontends | LiveKit Docs