Sentry

What it is

Sentry started as an open-source error-reporting tool — its signature feature is grouping thousands of occurrences of the same crash into a single “issue” with a count, instead of flooding a team with individual notifications. It has since grown into a full application observability platform: distributed tracing built on OpenTelemetry semantic conventions, a built-in continuous profiler, session replay, cron and uptime monitoring, and structured logs, all correlated against the same error and issue data that made Sentry popular in the first place. The self-hosted core remains open source, alongside a much more widely used commercial SaaS.

Why teams choose it

• Error grouping is still its strongest feature. Stack-trace-based issue grouping, with release tracking and regression detection, remains more mature than most general-purpose observability platforms’ approach to exceptions.
• One SDK covers errors, tracing, and profiling. Because tracing and profiling were added on top of the existing error SDK, adopting them from an existing Sentry install is usually a config change rather than a new integration.
• Strong default developer experience. Sentry is typically praised for how quickly a team can go from “add SDK” to “useful signal,” without needing to design dashboards first.
• Narrower than a full metrics/infrastructure platform. Sentry is application- and code-centric; teams still typically pair it with a dedicated metrics backend for infrastructure-level monitoring rather than replacing one with it.

Pricing model

Sentry is open source and self-hostable, but the overwhelming majority of usage is on its commercial SaaS, which offers a free tier for small projects and usage-based paid plans billed primarily on event and span volume across errors, tracing, and other signals.

Alternatives

Dash0 and Honeycomb are common alternatives for teams that want OpenTelemetry-native tracing without Sentry’s error-tracking-first design. Bugsnag and Rollbar are the closest direct competitors on pure error tracking. For profiling specifically, Grafana Pyroscope and Parca are open-source alternatives if profiling — rather than error tracking — is the primary need.