General
Observability
Overview
Monitor your application's performance and track errors.
The Pro Next.js Prisma starter kit includes structured logging and optional integrations for error, traffic and performance monitoring. Installing code is only the first step: Sentry and Vercel services must be enabled and verified in the environments where you expect them to collect data.
What works after setup
| Capability | Included in the repository | Required activation |
|---|---|---|
| Structured logs | Pino logger, grouped loggers and request context | Choose NEXT_PUBLIC_LOG_LEVEL and configure a log destination if needed |
| Sentry | Client, server and edge instrumentation plus build config | Add a Sentry project and DSN; add build credentials for source maps |
| Vercel Analytics | <Analytics /> in the root layout | Enable Web Analytics for the deployed Vercel project |
| Vercel Speed Insights | <SpeedInsights /> in the root layout | Enable Speed Insights for the deployed Vercel project |
Leaving a provider unconfigured should not stop the application from running, but that provider will not give you useful production telemetry.
Logging
Structured logging with Pino for high-performance log management.
Sentry
Error tracking and performance monitoring with Sentry.
Vercel Analytics
Real-time traffic data with Vercel Analytics.
Vercel Speed Insights
Real-time performance monitoring with Vercel Speed Insights.
Verify production monitoring
After deploying, prove that each enabled signal reaches its destination:
- Write a uniquely named structured log and find it in the deployment logs.
- Send a controlled test exception to Sentry and confirm that its stack trace resolves to the original source.
- Visit two or three production routes, then confirm page views appear in Vercel Analytics.
- Load a public page from a real browser and confirm Speed Insights begins collecting field data. This data may not appear immediately.
- Remove the test exception after verification and configure alerts for business-critical failures such as authentication, checkout and webhooks.
Keep environments separate
Use separate Sentry environments or projects for preview and production. Development noise and deliberate test failures should not trigger production incident alerts.