Sentry Prevent 🛡️ in BETA September 16 🚀

[rohan-at-sentry]

TL;DR

Sentry Prevent is a new set of capabilities that helps catch errors before they make it to production. When you introduce a new change, Sentry will pull existing Sentry issues (errors and performance) and try to predict if your changes will cause issues in production.

Sentry will also help identify flaky tests right in your Pull request saving developers time spent investigate test failures.

You can find onboarding instructions here

Just head to the Prevent 🛡️ tab on Sentry to get started.

Currently, we're only supporting Github Cloud, and Sentry orgs that are not based in the EU. We're working to add support to EU orgs actively and will keep this discussion updated as we iterate. For more details, keep reading or join the discussion below!

---

At Sentry, we know that it's much cheaper to find issues and critical bugs before we ship to production. This is why we write and execute automated tests as well as dogfood features internally before opening things up.

I'm sure most of you already know this to be true - and are likely using Code quality tooling in addition to testing. However, as users of these strategies ourselves, we've noticed that they have the following problems

1. Traditional code quality tools can be super noisy - providing feedback on code smells, code style / maintainability bugs etc at the same level as potentially serious bugs.
2. Testing infrastructure can take a long time to run (so it could slow down the time it takes to go from PR --> production)
3. With time, tests can become flaky. Flaky tests are institutional knowledge, and we've found that it really impacts velocity.

These problems effectively hinder your ability to ship higher quality code in the first place. This is why we're building Sentry Prevent 🛡️ .

For those of you who have been with us for some time, you know that we are super opinionated about .... well.. most things 😅 . And so, we approached Sentry Prevent with these design principles

1. Only highlight clear problems (errors, flaky tests) by default on the Pull request.
2. Provide feedback fast and quick, without blocking a developer.

Sentry Prevent helps solves the above problems by bringing 2 new capabilities to you

1. Prevent AI : which uses LLMs to provide targeted Pull Request feedback, specifically potential errors
2. Test Analytics : which uses test result reports you upload to us to determine if a test is flaky.

What can you do with Sentry Prevent?

• Get immediate feedback on your PR if you're introducing a potential error:

• Get immediate feedback on your PR if a failing test is known to be flaky

• Understand your test suite performance in Sentry: Find key metrics about your tests, organized by test suite - including run times, failure rates and the number of times your test have failed - across all branches, main, or any feature branch

Supported Integrations : Current and What's next

We're kicking things off with support for Github cloud and for Sentry Orgs not in EU. Broader EU support is coming over the next couple of months.

Try Sentry Prevent

Just head to the Prevent 🛡️ tab on Sentry to get started.

What Feedback Are We Looking For?

• How easy was it to get set up?
• Are there specific features or capabilities you’re missing?
• Is error prediction useful for your developers? Are we missing clear errors / highlighting the wrong things?

Your input will help us refine Sentry Prevent as we work towards general availability.

Drop your thoughts below, on Discord or email us directly at prevent-feedback@sentry.io. We’re looking forward to hearing from you (despite what that email address looks like 😅 )