AI-Driven Issue Detection - Early Access & Feedback #109234
mrduncan
announced in
Problem Understanding
Replies: 0 comments
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Uh oh!
There was an error while loading. Please reload this page.
-
TL;DR
We're working on AI-Driven Issue Detection - a new way of detecting problems in your application. Instead of relying only on hardcoded rules and pattern matching, we can now detect issues that don't match known strict patterns, use traces and logs together to identify problems, and surface higher-signal performance, security, and other types of issues. This is early, and we're looking for feedback.
Why We Built This
You've told us things like:
Traditional performance issue detection relies on deterministic detectors. This works well, but it has limits. It can't find problems outside the exact pattern it's written for, use broader trace context like logs, or adapt to new patterns without us shipping new detectors (which take time to build and tune).
AI-Driven Issue Detection analyzes traces and logs together to deliver application-specific insights by identifying performance and correctness problems (e.g. inefficient query patterns) as well as anti-patterns which emerge in runtime behavior.
This Is Early
We've seen it catch real issues that deterministic detectors missed. It also sometimes gets things wrong - and we want your help making it better.
You may see:
When that happens, tell us. Your feedback directly shapes what we fix and build next. We're actively iterating on quality, guardrails, and evaluation.
We'd Love Your Feedback
We're especially interested in:
If you're interested in actively testing this and giving feedback, email us at ai-issue-detection@sentry.io
Beta Was this translation helpful? Give feedback.
All reactions