Automatic Metrics

Every protect() call records execution metrics automatically. Use protect() and metrics appear on your Openfuse dashboard. No setup required.

What's collected

Window-based aggregation

Metrics are grouped into time windows (default: 10 seconds). Within each window, the SDK accumulates counts and collects latency samples.

Periodic flush

Every 15 seconds (default), the SDK sends completed windows to Openfuse in a single batch. In-progress windows are kept until they complete.

Reservoir sampling

To keep memory bounded, the SDK uses reservoir sampling to maintain a fixed number of latency samples per window (default: 1,000). This gives accurate percentile estimates at thousands of requests per second.

Instance deduplication

Each SDK instance has a unique ID to prevent double-counting. The SDK auto-detects your platform:

Override this with a custom instanceId in the client constructor.

Configure metrics

Default settings come from the server. Override them locally:

const openfuse = new OpenfuseCloud({
  system: 'payments',
  clientId: 'YOUR_CLIENT_ID',
  clientSecret: 'YOUR_CLIENT_SECRET',
  metrics: {
    windowSizeMs: 10_000,
    flushIntervalMs: 15_000,
    maxLatencySamples: 1_000,
    enabled: true,
  },
})

See Production for tuning advice.