---
title: "Performance Monitoring"
description: "Use traces to find latency, throughput, and regressions. Traces also show where work fails and how agent runs flow."
url: https://docs.sentry.io/product/sentry-basics/performance-monitoring/
---

# Performance Monitoring | Sentry Docs

Sentry goes beyond tracking page load metrics. By enabling [Tracing](https://docs.sentry.io/concepts/key-terms/tracing.md), Sentry automatically instruments the frameworks and libraries you use — capturing detailed performance data for every request across your entire application stack. This includes HTTP requests, database queries, template rendering, and more. Agent runs appear on the same trace when you set up [agent tracing](https://docs.sentry.io/product/agents.md) for an SDK that supports it. Some SDKs add them automatically, and others need manual setup.

## [Getting Set Up](https://docs.sentry.io/product/sentry-basics/performance-monitoring.md#getting-set-up)

Enable tracing in your SDK by setting `tracesSampleRate`. For most frameworks (Next.js, Laravel, Rails, Django, Express, and others), Sentry auto-instruments common operations like HTTP requests, database queries, and template rendering out of the box.

For applications with both frontend and backend services, set up [distributed tracing](https://docs.sentry.io/concepts/key-terms/tracing/distributed-tracing.md) to follow requests across service boundaries. Learn more about [trace propagation](https://docs.sentry.io/platforms/javascript/tracing/distributed-tracing.md) for your platform.

## [Ways to Monitor Performance](https://docs.sentry.io/product/sentry-basics/performance-monitoring.md#ways-to-monitor-performance)

### [Pre-built Dashboards](https://docs.sentry.io/product/sentry-basics/performance-monitoring.md#pre-built-dashboards)

After turning on tracing, you'll automatically see [Sentry Dashboards](https://docs.sentry.io/product/dashboards/sentry-dashboards.md) that give you a high-level view of performance data for each layer of your stack:

* [**Frontend**](https://docs.sentry.io/product/dashboards/sentry-dashboards/frontend.md) — Core Web Vitals, transaction duration, error rates
* [**Backend**](https://docs.sentry.io/product/dashboards/sentry-dashboards/backend.md) — database queries, API calls, caches, queues, crons
* [**Mobile**](https://docs.sentry.io/product/dashboards/sentry-dashboards/mobile.md) — cold/warm app starts, TTID & TTFD, screen load times
* [**AI**](https://docs.sentry.io/product/agents.md) — LLM call latency, token usage, tool execution performance



### [Automated Performance Issues](https://docs.sentry.io/product/sentry-basics/performance-monitoring.md#automated-performance-issues)

Sentry automatically detects common performance problems and groups them into [performance issues](https://docs.sentry.io/product/issues/issue-details/performance-issues.md):

* N+1 queries and API calls
* Consecutive DB queries or HTTP requests that could be batched
* Database or file I/O on the main thread (blocking UI)
* [And more](https://docs.sentry.io/product/issues/issue-details/performance-issues.md)

These appear alongside error issues in your [Issues](https://docs.sentry.io/product/issues.md) feed, so you can triage and resolve them the same way.

### [Trace Explorer](https://docs.sentry.io/product/sentry-basics/performance-monitoring.md#trace-explorer)

[Trace Explorer](https://docs.sentry.io/product/trace-explorer.md) queries span data by any attribute. Find the p95 duration of a specific endpoint, identify which user segments are experiencing slowdowns, or compare a time window against your baseline to isolate regressions.

You can also group span durations by [custom span attributes](https://docs.sentry.io/product/trace-explorer.md#aggregating-span-attributes) to compare operations, such as p95 checkout duration by payment provider, and use these queries in [dashboards](https://docs.sentry.io/product/dashboards.md) and [monitors](https://docs.sentry.io/product/monitors-and-alerts/monitors.md). Page filters and saved queries are covered in [Explore](https://docs.sentry.io/product/explore.md).



### [Connected Context](https://docs.sentry.io/product/sentry-basics/performance-monitoring.md#connected-context)

Performance data in Sentry isn't siloed. Every trace connects to:

* **[Logs](https://docs.sentry.io/product/logs.md)** — see what your application logged during a slow request
* **[Session Replay](https://docs.sentry.io/product/session-replay.md)** — watch the user experience during a performance issue
* **[Profiling](https://docs.sentry.io/product/profiling.md)** — drill from a slow span into the exact functions consuming CPU time
* **[Agent Tracing](https://docs.sentry.io/product/agents.md)** — trace agent workflows, LLM calls, and tool executions alongside your application traces
* **[Seer](https://docs.sentry.io/product/ai-in-sentry/seer.md)** — get AI-powered root cause analysis for performance regressions

## [How It Works Under the Hood](https://docs.sentry.io/product/sentry-basics/performance-monitoring.md#how-it-works-under-the-hood)

Instead of capturing predefined metrics at fixed intervals, Sentry instruments the libraries you rely on, providing detailed telemetry for Web Vitals, file I/O, network requests, and more. Data is recorded as [spans](https://docs.sentry.io/concepts/key-terms/tracing.md#whats-a-span), the steps inside a [trace](https://docs.sentry.io/concepts/key-terms/tracing.md). Traces are assembled server-side. The tools on this page aggregate them, and you can still open the trace to see where the work failed or slowed down.

## [Next Steps](https://docs.sentry.io/product/sentry-basics/performance-monitoring.md#next-steps)

* **[Querying Traces guide](https://docs.sentry.io/get-started/guides/querying-traces.md)** — practical queries for finding slow pages and bottlenecks
* **[Adding Custom Spans guide](https://docs.sentry.io/get-started/guides/custom-spans.md)** — instrument your own code for visibility beyond auto-instrumentation
* **[Getting Started tutorial](https://docs.sentry.io/product/sentry-basics/getting-started-tutorial.md)** — follow along with a sample React + Express app
