BugWatch
BugWatch is the New Instance observability product: error monitoring, structured logs, distributed traces, metrics and release health for every tier of your stack - backend services, browsers, and native mobile apps - in one place.
What BugWatch doeslink
When your application throws, crashes, or logs, BugWatch captures the event with its full context (stack trace, user, tags, release, breadcrumbs, trace correlation) and turns the stream into something actionable:
- Issues - events are grouped by fingerprint into issues, each with first/last-seen, occurrence counts, affected users, level and status (open → resolved/ignored). One noisy bug is one issue, not ten thousand rows.
- Logs - structured log search across every service and environment, with live tail in the dashboard.
- Traces & metrics - OpenTelemetry-native: OTLP ingest, Prometheus-compatible queries, Jaeger-compatible trace browsing, a service map, SLOs, and RUM web-vitals from real browsers.
- Release health - sessions and crash-free rates per release, so a bad deploy is visible in minutes.
- Symbolication - minified JS, R8/ProGuard-obfuscated JVM, Dart-obfuscated and native (dSYM/ELF) stack traces are resolved server-side from artifacts you upload in CI.
- Deploy correlation - CI-reported deployments show up on a timeline next to error rates; a failed deploy opens an issue automatically.
- Alerting & AI root cause - threshold and issue-count alert rules notify your webhook; one-click AI root-cause analysis summarises what broke and why (also exposed to AI agents via AI Agent Access (MCP)).
Core conceptslink
| Concept | What it is |
|---|---|
| Project | One monitored application. Created in the dashboard; owns events, issues, symbols, deployments and settings. |
| DSN key | The project's server credential - a central API key (sk_…KEYID:secret) scoped to that project. Used by server SDKs, CI tooling and direct API calls. |
| Mobile credentials | projectId + appSecret for mobile apps - the secret never leaves the device; batches are signed with 5-minute HMAC tokens. |
| Session token | Short-lived browser credential minted by your backend, so the secret never reaches the client. |
| Environment | A per-event label (production / staging / development) bound to the key's scope - one project, separated streams. |
| Release | Your version string. It ties events to symbol uploads, release health and deploys - keep it identical across SDK init, symbol upload and deploy reporting. |
How this section is organisedlink
Work top to bottom - it follows the integration journey:
- Install an SDK - server & web packages (JS/TS, PHP), the CLI, the OTel collector, and the raw-REST path for everything else.
- Mobile - dedicated per-platform guides: Android, iOS & macOS, React Native, Flutter.
- SDK Usage & Event Enrichment - users, tags, context, releases, sampling, redaction, dedup, trace correlation.
- Browser Ingest - the session-token flow for browser apps and the browser wire format.
- Server Ingest API - the wire reference: exactly what goes over HTTPS.
- OpenTelemetry, Prometheus & Tracing - OTLP in, Prometheus/Jaeger-compatible reads, RUM.
- Source Maps & Symbols - CI uploads that make production stack traces readable.
- Deployments (CI/CD) - report deploys so errors correlate with releases.
Before you start: create a project (Dashboard → Org → BugWatch → New project - the guided wizard generates a setup recipe for your exact language and framework) and copy its DSN key. Every request in this section authenticates with it (x-api-key: KEYID:secret) unless the page says otherwise; the scope each endpoint needs is listed on the request, and the full scope catalogue lives in Getting Started.