Back to Blog Home

How we sped up Metric Alerts

Chris Heher image

Chris Heher -

Dogfooding Chronicles: Thinking Backward, Moving Forward

Zac Propersi, Engineering Manager at Sentry, can tell when a page is not loading as fast as it should — just by looking at it. While working on our new Metric Alerts feature, Zac noticed that the alerts pages were rendering slowly. Being the super Sentry user that he is, he wrote a custom query in Discover to see just how slow the transactions were.

Reproduction

By using the tracing capabilities as a part of our new performance monitoring offering, Zac found out that an incident’s endpoint call was taking seven seconds to complete. This, along with a handful of other queries, were powering the trend graphs and alert statistics which impacted load time.

Resolution

Enter Billy Vong and Dan Fuller, specialists in delivering quality customer experiences. They solved the problem by deferring the slow-loading queries to run shortly after the page loads. This way, the most relevant alert data would load first and then trend graphs and supporting statistics would follow. The result? A landing page that loads in less than 300 ms.

Application monitoring means monitoring all aspects of your code, not just your errors. From getting notified with Metric Alerts and analyzing the code with Discover, to using tracing the issue with Performance, the team was able to run circles around the frontend/backend feedback loop.

Share

Share on Twitter
Share on Facebook
Share on HackerNews
Share on LinkedIn

Published

Sentry Sign Up CTA

Code breaks, fix it faster

Sign up for Sentry and monitor your application in minutes.

Try Sentry Free

Topics

Dogfooding Chronicles

The best way to debug slow web pages

Listen to the Syntax Podcast

Of course we sponsor a developer podcast. Check it out on your favorite listening platform.

Listen To Syntax
    TwitterGitHubDribbbleLinkedinDiscord
© 2024 • Sentry is a registered Trademark of Functional Software, Inc.