Diagnostic tool for WebRTC JS applications that analyzes WebRTC getStats() result in realtime and generates a report on possible issues.
- Mean opinion score - calculates MOS for inbound and outbound network connections that can indicate a problem before it even appears.
- CPU issues - indicates possible issues with encoding and decoding media streams.
- Server issues - indicates possible server side issues.
- Fully customizable - allows to create your own detectors or WebRTC getStats() parsers.
yarn add webrtc-issue-detector
import WebRTCIssueDetector from 'webrtc-issue-detector';
// create it before the first instance of RTCPeerConnection is created
const webRtcIssueDetector = new WebRTCIssueDetector({
onIssues: (issues) => issues.map((issue) => {
console.log('Issues type:', issue.type); // eg. "network"
console.log('Issues reason:', issue.reason); // eg. "outbound-network-throughput"
console.log('Stats:', issue.statsSample); // eg. "packetLossPct: 12%, avgJitter: 230, rtt: 150"
}),
onNetworkScoresUpdated: (scores) => {
console.log('Inbound network score', scores.inbound); // eg. 3.7
console.log('Outbound network score', scores.outbound); // eg. 4.5
console.log('Network stats', scores.statsSamples); // eg. { inboundStatsSample: { avgJitter: 0.1, rtt: 30, packetsLoss: 8 }, ... }
}
});
// start collecting getStats() and detecting issues
webRtcIssueDetector.watchNewPeerConnections();
// stop collecting WebRTC stats and issues detection
webRtcIssueDetector.stopWatchingNewPeerConnections();By default, WebRTCIssueDetector can be created with minimum of mandatory constructor parameters. But it's possible to override most of them.
import WebRTCIssueDetector, {
QualityLimitationsIssueDetector,
FramesDroppedIssueDetector,
FramesEncodedSentIssueDetector,
InboundNetworkIssueDetector,
OutboundNetworkIssueDetector,
NetworkMediaSyncIssueDetector,
AvailableOutgoingBitrateIssueDetector,
UnknownVideoDecoderImplementationDetector,
} from 'webrtc-issue-detector';
const widWithDefaultConstructorArgs = new WebRTCIssueDetector();
// or you can fully customize WebRTCIssueDetector with constructor arguments
const widWithCustomConstructorArgs = new WebRTCIssueDetector({
detectors: [ // you are free to change the detectors list according to your needs
new QualityLimitationsIssueDetector(),
new FramesDroppedIssueDetector(),
new FramesEncodedSentIssueDetector(),
new InboundNetworkIssueDetector(),
new OutboundNetworkIssueDetector(),
new NetworkMediaSyncIssueDetector(),
new AvailableOutgoingBitrateIssueDetector(),
new UnknownVideoDecoderImplementationDetector(),
],
getStatsInterval: 10_000, // set custom stats parsing interval
onIssues: (payload: IssueDetectorResult) => {
// your custom callback for detected issues handling
},
onNetworkScoresUpdated: (payload: NetworkScores) => {
// your custom callback for networks score updates handling
},
ignoreSSRCList: [
// in case you need to skip some ssrc from parsing, add its numbers to the array
],
});Detects issues with outgoing network connection.
const exampleIssue = {
type: 'network',
reason: 'outbound-network-throughput',
statsSample: {
availableOutgoingBitrate: 1234,
videoStreamsTotalBitrate: 1234,
audioStreamsTotalTargetBitrate: 1234,
},
}Detects issues with decoder.
const exampleIssue = {
type: 'cpu',
reason: 'decoder-cpu-throttling',
statsSample: {
deltaFramesDropped: 100,
deltaFramesReceived: 1000,
deltaFramesDecoded: 900,
framesDroppedPct: 10,
},
ssrc: 1234,
}Detects issues with outbound network throughput.
const exampleIssue = {
type: 'network',
reason: 'outbound-network-throughput',
statsSample: {
deltaFramesSent: 900,
deltaFramesEncoded: 1000,
missedFramesPct: 10,
},
ssrc: 1234,
}Detects issues with inbound network connection.
const exampleIssue = {
type: 'network',
reason: 'inbound-network-quality' | 'inbound-network-media-latency' | 'network-media-sync-failure',
iceCandidate: 'ice-candidate-id',
statsSample: {
rtt: 1234,
packetLossPct: 1234,
avgJitter: 1234,
avgJitterBufferDelay: 1234,
},
}Also can detect server side issues if there is high RTT and jitter is ok.
const exampleIssue = {
type: 'server',
reason: 'server-issue',
iceCandidate: 'ice-candidate-id',
statsSample: {
rtt: 1234,
packetLossPct: 1234,
avgJitter: 1234,
avgJitterBufferDelay: 1234,
},
}Detects issues with audio synchronization.
const exampleIssue = {
type: 'network',
reason: 'network-media-sync-failure',
ssrc: 1234,
statsSample: {
correctedSamplesPct: 15,
},
}Detects issues with outbound network connection.
const exampleIssue = {
type: 'network',
reason: 'outbound-network-quality' | 'outbound-network-media-latency',
iceCandidate: 'ice-candidate-id',
statsSample: {
rtt: 1234,
avgJitter: 1234,
packetLossPct: 1234,
},
}Detects issues with encoder and outbound network. Based on native qualityLimitationReason.
const exampleIssue = {
type: 'cpu',
reason: 'encoder-cpu-throttling',
ssrc: 1234,
statsSample: {
qualityLimitationReason: 'cpu',
},
}const exampleIssue = {
type: 'network',
reason: 'outbound-network-throughput',
ssrc: 1234,
statsSample: {
qualityLimitationReason: 'bandwidth',
},
}Detects issues with decoding stream.
const exampleIssue = {
type: 'stream',
reason: 'unknown-video-decoder',
ssrc: 1234,
trackIdentifier: 'some-track-id',
statsSample: {
mimeType: 'video/vp9',
decoderImplementation: 'unknown'
},
}- Adaptive getStats() call interval based on last getStats() execution time
- Structured issue debug
- Issues detector for user devices permissions
Pull requests are welcome. For major changes, please open an issue first to discuss what you would like to change.
Please make sure to update tests as appropriate.