James McDowell - Reducing incident response time through targeted App Observability & Alerting скачать в хорошем качестве

James McDowell - Reducing incident response time through targeted App Observability & Alerting

Reducing incident response time through targeted App Observability & Alerting by James McDowell (Senior Engineer - Mobile @ REA Group) Delivered at GDG Melbourne Devfest 2025 One of the most difficult things to address is vague user complaints, usually through app reviews - but in the world of mobile apps, this is classically the main form of feedback that you see about your apps in production. If you're lucky, the reviewer has given you enough information for you to start mucking around in your app. If you're REALLY lucky, they've given you some form of reproduction steps. If you're unlucky though, and you can't reproduce the problem locally, what do you do? There's so many devices running Android; do you have the time and resources to test them all? Probably not, so you'll leave it until someone else reports the problem. But what if it didn't have to be this way? In the world of backend, it's normal and expected for systems to be integrated into a tool such as New Relic which reports on their health. Mobile apps operate differently though, can we really reach the same level of rich information? Is it really worth the effort? In this talk, I'll show you how we drove this change at REA-Group, how it shows up in our day-to-day operations, and the value it's driving for our business. We'll dive into the ongoing observability metrics we focus on, and walk through an example of how our app's real-time alerting has contributed to ongoing incidents both within and outside of the mobile space. Finally, I'll touch on where we're going next and how that direction will improve our team velocity.