In local monitoring, the endpoint fetches and analyses the local metrics, i.e., metrics related to the incoming or outgoing media stream. Ergo a stream at Endpoint A only knows the local performance of the incoming and outgoing streams and not the performance of the those streams at a remote endpoint.
Alternatively, an end-to-end monitoring solution monitors the performance of the streams at each endpoint, it then sends the data to a centralised server from each endpoint in a call. The centralised server can then correlate the performance of the media streams across endpoints. Therefore, using this process allows the centralised server to diagnose performance issues.
callstats.io is a scalable centralised server that monitors receives and analyses the measurement data (metrics) from each endpoint in a call in real-time. We aggregate measurement data at 3 levels:
- at each endpoint in a call (local monitoring)
- at the conference level, i.e., to detect and diagnose issues by correlating across each endpoint.
- at the service level, i.e., call failures, call drops, etc.