Sentry

Sentry is application monitoring and error tracking software for developers that links errors, traces, logs, session replays and profiles, billed by usage.

Sentry shows developers which line of code broke in production, what the user was doing and which request carried the failure. Functional Software, Inc., a San Francisco company that trades as Sentry, owns and builds it. It works across frontend, mobile and backend code, in any language or framework. Errors, traces, logs and user sessions are tied together, so one issue holds the evidence a fix needs. Error Monitoring groups exceptions into issues with code-level context, triage tools and Release Health for each deploy. Tracing follows a request from the first user action through every service and function call, drawn as a waterfall in Trace View. Profiling watches live backend and mobile code continuously to find the slow function or line, while Query Insights ranks slow database queries. Logs keeps structured log lines beside the trace that emitted them, and Session Replay plays back a user's session with sensitive data masked. Seer, the debugging agent, investigates a new issue on its own and drafts a fix as a pull request. Sentry suits developers and engineering teams who own their code after it ships, from one person on a side project to organizations that manage access through SAML and SCIM. Frontend teams use replays to see what a user saw before an error, and mobile teams follow crashes and release health. Backend teams lean on traces and profiles to explain a slow endpoint or a heavy query. Teams that keep separate tools for errors, logs and traces gain the most, because one Sentry issue links all three. A typical issue starts as an alert from a Metric Monitor, opens on a stack trace and leads to the replay, the trace and the logs around it. Seer may already have a pull request waiting in GitHub, and AI code review checks the next change against real error history before it merges. Uptime Monitoring and Cron Monitoring cover the endpoints and scheduled jobs that fail quietly. Teams instrumented with OpenTelemetry send traces and logs over OTLP instead of swapping SDKs. Functions on AWS Lambda, Azure Functions and Google Cloud Functions are traced inside the request that called them, so a slow function call shows up where the user felt it.