# Sentry
Sentry is an error tracking and application performance monitoring platform used by developers to diagnose, fix, and optimize the runtime behavior of their code. Originally an open-source project, it has grown into a 200-person company processing 790 billion events per month for 4 million developers across 150,000 organizations.
## What It Does
Sentry captures, deduplicates, and contextualizes errors and performance issues across web, mobile, and backend stacks — turning raw stack traces into actionable issues with breadcrumbs, request data, releases, commits, and user impact. The platform spans:
- **Error monitoring** — the original product
- **Performance monitoring & tracing** — distributed traces, slow transactions
- **Profiling** — sampled call-stack data
- **Logs**
- **Session Replay** — DOM-level recording of user sessions
- **Cron monitoring** & **Uptime monitoring**
Supports 100+ languages and integrates with GitHub, GitLab, Slack, Jira, and most CI/CD systems.
## AI-Native Features
A new layer of AI products sits on top of the data Sentry already collects:
- **Seer** — agentic issue analysis and auto-debugging
- **Autofix** — proposes patches for tracked errors
- **AI Code Review** — uses [[Warden]], Sentry's open-source agentic code review tool driven by [[AI Agent Skills]]
- **MCP Server** — gives AI agents structured access to Sentry issues for terminal-based querying
This positioning is notable: Sentry has the runtime error data; layering agents on top of it is a natural extension into [[DevSecOps]] and developer productivity.
## History
- Started as an open-source project by founders previously at Dropbox, GitHub, and Atlassian, including [[David Cramer]]
- Grown into the leading error-tracking platform
- Raised $217M across 6 rounds
- Offices in San Francisco, Seattle, Vienna, Toronto
- Pioneered the [[Functional Source License (FSL)]], a Fair Source license now used across Sentry projects including [[Warden]]
## Why It Matters Beyond Error Tracking
Sentry sits at an interesting position in the developer toolchain: it sees what actually breaks in production. That makes its data uniquely useful as input to AI agents — Autofix isn't reasoning in a vacuum, it has the real stack trace, the real release diff, and the real user context. Most AI coding tools have to *infer* what's wrong; Sentry agents *know*.
## References
- https://sentry.io
- https://sentry.io/about/
- https://docs.sentry.io
- https://github.com/getsentry/sentry
## Related
- [[David Cramer]]
- [[Warden]]
- [[Warden CLI]]
- [[Functional Source License (FSL)]]
- [[AI Observability]]
- [[DevSecOps]]
- [[GitHub]]