Raven Docs

Reconnection & Network Quality

What happens when the network drops, and how much control you have over it — on every SDK.

The SDK surfaces Raven's own reconnect policy — exponential backoff with a capped retry delay, then a clean failed state — rather than reimplementing reconnect logic on top of it. The policy is identical across platforms because it's the same underlying client.

createRTCClient({ token, endpoint, autoReconnect: true }); // default
  • autoReconnect: true (default): automatic reconnect on network loss, surfaced as reconnecting → reconnected — or disconnected plus an error event with code CONNECTION_FAILED if every retry is exhausted.
  • autoReconnect: false: the first disconnect goes straight to disconnected. No retries.
room.on('reconnecting', () => showBanner('Reconnecting…'));
room.on('reconnected', () => hideBanner());
room.on('error', (error) => {
  if (error.code === 'CONNECTION_FAILED') {
    showError('Lost the connection. Try rejoining.');
  }
});

What a reconnect looks like to a participant

Tracks stay subscribed through a brief reconnect — you generally don't need to re-attach media elements. A disconnected → failed transition is the signal that the room needs to be rejoined from scratch, not merely reconnected.

App lifecycle (React Native)

The SDK watches AppState and reports transitions, but deliberately does not disconnect on background — dropping the socket on a brief app-switch would turn "checked a notification" into "left the meeting." Video capture does stop when backgrounded (the OS suspends the camera) and resumes on return.

new Raven({
  onAppStateChange: (state) => {
    if (state === 'background') void raven.leave(); // your call, not automatic
  },
});

See Background Audio for keeping the call alive (audio-only) while backgrounded, instead of leaving it.

Watching quality in real time

room.getConnectionStats() gives you the numbers behind "how's this call actually going" — see Diagnostics. A common pattern is polling it to drive a quality indicator in your own UI:

setInterval(async () => {
  const stats = await room.getConnectionStats();
  updateQualityBadge(stats.connectionQuality); // 'excellent' | 'good' | 'poor' | 'lost' | 'unknown'
}, 5000);

This works identically on Web and React Native (same Room class).

If you're on the dashboard side rather than in the client, the same numbers are already flowing into Raven's own telemetry — see Event Catalogue and raven connections inspect.

Common errors

ErrorWhyFix
CONNECTION_FAILEDEvery reconnect attempt was exhausted.Rejoin from scratch — client.join()/raven.join() again with a fresh token if the old one has expired.
Reconnect never firesautoReconnect: false was set.Remove it, or handle disconnected yourself and call join() again.

Production notes

  • Don't build your own retry loop around join() — you'd be racing the SDK's own reconnect logic. Listen for failed and rejoin then, not before.
  • A reconnecting banner that never clears usually means the app crashed mid-reconnect, not that the SDK is stuck — check for an uncaught error nearby.