Quick Start

Get up and running with GoodLogs in under 2 minutes. No infrastructure to set up — just install the SDK and start tracking.

1. Install the SDK

bash

npm install @aj-2000-test/goodlogs-sdk

2. Initialize

typescript

import { GoodLogs } from "@aj-2000-test/goodlogs-sdk"

const gl = new GoodLogs({
  apiKey: "gl_pk_your_key_here",
  // Opt into the unified APM pipeline (errors, traces, web vitals).
  useEnvelope: true,
})

Get your API key from Dashboard → Settings → API Keys.

With useEnvelope: true the browser SDK automatically:

wraps every fetch() call as a span (op:http.client) and injects a W3C traceparent header,
collects Core Web Vitals (LCP / INP / CLS / FCP / TTFB) via PerformanceObserver,
captures window.onerror, unhandled rejections, and console errors as Issues.

3. Track Events

typescript

gl.track("signup", { plan: "pro", source: "landing_page" })

4. Send Logs

typescript

gl.error("Payment failed", {
  orderId: "ord_123",
  error: "card_declined",
  service: "billing",
})

5. Capture Exceptions

typescript

try {
  await chargeCard()
} catch (e) {
  gl.captureException(e, { tags: { feature: "checkout" } })
}

Errors are fingerprinted and grouped into Issues in the dashboard. Duplicates bump the event count instead of creating new issues. If you resolve an issue and the same error appears on a later release, it's automatically flipped to regressed.

6. Trace a slow flow

typescript

const tx = gl.startTransaction({ name: "/checkout", op: "ui.action" })
const span = tx.startChild({ op: "http.client", description: "POST /pay" })
await fetch("/api/pay", { method: "POST" })
span.finish()
tx.finish()

Open Dashboard → Performance for endpoint latency (p50/p95/p99) and slow traces. Click a trace to see the waterfall.

7. View in Dashboard

Page	What you'll see
Issues	Grouped exceptions with stack traces, breadcrumbs, and trace links
Performance	Endpoints sorted by p95, Core Web Vitals tiles, recent slow traces
Trace (`/dashboard/traces/<id>`)	Span waterfall with attributes and events
Web Vitals	LCP / INP / CLS rating distributions, slowest pages per metric
Session Replay	DOM recordings of user sessions with linked console/network/errors
Profiling	CPU flamegraph, top functions by self-time, profile chunk list
Explorer	Ad-hoc GQL across logs, events, errors, spans, web_vitals
AI Debug	Plain-English questions answered with GQL

Tip: The SDK auto-batches and flushes every 5 seconds. On page close, it uses navigator.sendBeacon to ensure no data is lost.

Zero-config path: Even without captureException / startTransaction calls in your code, GoodLogs background workers automatically:

synthesise Issues from any log at severity:error or severity:fatal,

synthesise traces from logs that share a request_id / correlation_id property — useful when migrating from plain logging.