Release Health is an early warning system that mobile and native application developers rely on to quickly assess and course-correct a bad release. The first 5 to 10 minutes after code is pushed to production are arguably some of the most stressful and uncertain minutes for a dev team — you can’t shake that did-anything-break?! feeling. As Popsa CTO, Tom Cohen, put it, “Release Health is a game-changer for us. Being able to detect bad releases quickly and track crash-free users over time helps us move faster as a team and lets us solve what matters most to our customers.”
# Using yarn yarn add @sentry/browser @sentry/tracing # Using npm npm install --save @sentry/browser @sentry/tracing
Here’s how we define Sessions:
- Session - A page view
- New Session - Every refresh
- Multiple Sessions - Multiple browser tabs
- Crash Free Sessions: Number of sessions that did not end by the crash of the application
- Crash Free Users: Percentage of the users that did not experience a crash
- Version adoption: Number of users you probably need to nudge to refresh their browser
- Number of new issues introduced in the latest release
With Release Health configured by default, you can track how a release trends from the second it’s live to the point it starts to degrade (a rare moment, I know.).
Jaylum Chen, Staff Software Engineer at Eventbrite, and his team have been using Release Health, to quickly detect bad releases by gauging the levels of adoption and the percentage of crash-free sessions and crash-free users. Sentry takes the mystery out of debugging and gives the Eventbrite engineering team the context needed to quickly fix an issue. This lets them focus their time on building what’s next rather than debugging hard to find bugs.
Sentry helps our team fix the most important issues in each release. We can zoom in on the issues that impact the most users and get it resolved so we can move on to building more features.
Eventbrite sees Release Health as integral in being able to achieve CI/CD to support an increasing number of events. Migrating from their “Core” monolith into microservices also means higher complexity in troubleshooting. Sentry helps the infrastructure and feature teams easily isolate issues by helping developers gauge performance, adoption, and usage — accelerating Eventbrite’s journey to continuous deployment.
Coupled with Discover, Sentry’s insights tool, teams can get to the root cause of any issue. Logins drop after a release? Support tickets from a specific region spiking? Click the “Open in Discover” button from any release details page and Sentry will automatically build and run a query for events by the release.
Analyze data by URL, geography, device, browser, etc. to see new issues by release, solve the ones that are impacting your users and business as well as learn how your releases are trending over time
This update also comes with support for WebAssembly. If you have a compiler that can produce DWARF debug data you can now get crash reports including source code on the server. Don’t know how big of a deal this is, read about our gripes with WASM before this release.