There’s never a good time for errors or performance degradations to show up, but during periods of peak traffic like the holidays, it’s especially critical to get immediate answers about what's failing and how to fix it.
If you’ve set up a Sentry metric alert to alert on the number of errors in your application, you’ve likely found that this count of errors includes events from archived, resolved, and unresolved issues. This can be less than ideal, since spiking archived issues can trigger your metric alert and create unnecessary noise.
Previously, the best solution was to exclude noisy archived issues manually via !issue.id:123 which would exclude events from that issue in your metric alert. You can now use is:unresolved directly in metric alerts which automatically excludes events from archived or resolved issues.
We’ve also added banners to your existing error metric alerts if they’re not using is:unresolved. When applying this new filter to your metric alert, you should review the threshold values for critical, warning, and resolved since the total number events will have changed.
Github
Slack
Javascript
Python
React
Laravel
Next.js