Rustrak

What Rustrak does

Rustrak is an error tracker that implements the Sentry ingestion protocol. Instead of shipping its own client libraries, it accepts the envelopes that the official Sentry SDKs already send, so an application in Python, JavaScript, Go or any other language with a Sentry SDK reports to Rustrak once its DSN points at a Rustrak project. The project suggests adding Rustrak as a second DSN next to an existing Sentry setup to compare the two before switching.

Events are grouped into issues by a fixed rule: the fingerprint the SDK sent, if any, otherwise the exception type, the first line of the message and the transaction name. The aim is that the same error always lands in the same issue, including after a deploy. The issue list supports bulk triage and status filters and shows a 24-hour trend per issue, and uploaded source maps turn minified JavaScript frames back into source lines.

Beyond errors, Rustrak takes the other data types Sentry SDKs emit: transactions and spans for a performance view with span waterfalls, structured logs with six severity levels that link to the trace they came from, sessions for per-release crash-free rates, and traces from the SDKs’ AI integrations showing model calls, tool calls and token counts.

Alerts and access

Alert rules fire when a new issue appears, when a resolved issue comes back, or when a muted issue is unmuted, and each rule sends to Slack, SMTP email or a JSON webhook. Projects have Admin, Member and Viewer roles with invitations. Password login can be combined with OpenID Connect single sign-on through providers such as Keycloak, Authentik, Pocket ID, Google or Microsoft Entra. A REST API with a typed TypeScript client and an MCP endpoint let scripts and AI assistants query issues.

Self-hosting Rustrak

The default install is one Docker Compose file and one container that serves both the API and the dashboard, storing everything in SQLite on a volume. You set a session secret, an initial admin account and the public URL that SDKs will send to. The documentation says the server idles under 100 MB of memory and describes SQLite as suitable for personal projects, small teams and up to about 1,000 events an hour; above that, a separate image uses PostgreSQL. The dashboard can also run on its own host.

Retention is configurable, and the server reports its own disk usage, with manual cleanup and source map garbage collection. By default the server sends an anonymous report every six hours with version, platform, memory and request counters; setting RUSTRAK_TELEMETRY=off or DO_NOT_TRACK=1 turns it off.

Hosted version

There is none. The README states there is no paid tier and no hosted plan.

Limitations

Rustrak is young, still below version 1.0, and is copyrighted to and funded through the sponsorship of a single developer, which matters for a tool you rely on during incidents. The documentation lists attachments, session replays and profiling as unsupported, so Sentry users who depend on those will lose them. You carry the operational work yourself: backups, upgrades and, above low volumes, a PostgreSQL database.

Who it suits

Rustrak suits small teams and individual developers who already use Sentry SDKs and want a self-hosted server that fits on an existing machine without Kafka or ClickHouse. Teams that need replay or profiling, or want a hosted service, should compare the options in the error tracking guide and on the Sentry alternatives page.