Diagnostics

Observe Insight, vf ingest, and WebRTC delivery. Start with the compact views for active channels, then add all=1&verbose=1 when diagnosing inactive ports, packet loss, malformed RTP, missing keyframes, RTCP feedback, or browser decode problems.

📄️Get a system metrics snapshot

Returns the latest point-in-time metrics assembled for the Insight dashboard, including CPU, memory, disk, temperature, MLA, remote-platform, and pipeline compatibility fields. The payload is intentionally extensible, so clients should tolerate additional fields.

📄️Stream Neat metrics events

Opens a long-lived server-sent events connection to the local metrics broker. Each SSE `data` field contains one JSON metrics event; clients should reconnect after network interruption and must not wait for the response body to terminate.

📄️Get vf UDP and RTP ingest statistics

Shows whether video RTP and metadata JSON are reaching vf before WebRTC delivery is considered. A healthy H.264 input normally has increasing packet and bitrate counters, observed SPS/PPS, and periodic IDR frames. Add `verbose=1` for payload types, NAL counts, sequence gaps, jitter estimates, malformed packets, and recent errors; add `all=1` to include inactive channels.

📄️Get vf WebRTC egress statistics

Use this after ingest statistics confirm that RTP is arriving but a browser does not decode, render, or remain connected. The response combines vf peer state, RTCP feedback, metadata DataChannel counters, and the latest browser `RTCPeerConnection.getStats()` report. Add `verbose=1` for recent peer errors and `all=1` to retain inactive peers.