Making Sentry notifications smart(er) with ML

[rachrwang]

Hi everyone!

I’m Rachel from Sentry’s product team, here with @trillville and @pevensentry to share an exciting update on the future of notifications at Sentry.

Over the last few months, we’ve been developing ML capabilities that allow us to assess whether an issue is likely to be important for surfacing in real-time based on prior user interactions with similar issues. Our first feature is the ability to alert on new issues that pass a certain severity threshold on its first event, or fires with high frequency shortly after creation.

We'd love to hear from you on the following questions:

1. Would this new alert option be valuable for you and your team?
2. What is your typical workflow for triaging Sentry errors? How do you identify which errors to care or not care about?
3. If you had a magic wand to improve how your team discovers new issues, what would you want to see?

Want to participate in our upcoming beta program? All we need is for the owner of your org to sign this acknowledgement, which allows our self-hosted models to use your event data for this feature.