Evaluation and observability

Sentry

Error tracking with the stack trace and the release.

made by
Sentry

Magic Ship is one shop in Vancouver, BC, working remotely with clients worldwide. We are not a partner, reseller, or certified vendor of Sentry - we just build with this.

What Sentry is

Sentry captures exceptions and performance data from backend and frontend code, groups them into issues, and ties each one to a release and a commit through source maps and release tracking. It traces across services and can be self-hosted.

How we use it

Projects we hand over report errors here with the release version attached, so "when did this start" has an answer instead of a theory. Frontend source maps are uploaded at build time, which is the difference between a readable stack trace and a minified one. Alert rules are kept few enough that people still read them.

Where it is the wrong choice

It catches exceptions, not wrong answers. An agent that returns a confident, incorrect response throws nothing, so Sentry stays quiet while the product is broken - that failure mode belongs to evals and traces.

Building something on Sentry?

Send the problem rather than a job spec. You get an answer on scope, on fit, and on whetherSentry is even the right call for it.

Start a project

Sentry and Sentry are trademarks of their respective owners, used here to say what we work with.