Are you smart enough to predict an outage?

All cloud providers experience outages, that can throw a wrench into apps. Recent events, such as the AWS outage in the Middle East, have once again highlighted the unpredictable nature of technology and cloud services.  No app is bulletproof, any app can be knocked offline by an unexpected outage. One Shorebird customer was able to quickly recover by pushing patches to implement fallback logic and redirect requests to secondary services. 

This shines a light on a common problem that mobile devs face. Web devs can push code at a moments notice, this isn’t a luxury that mobile devs have based on app store policies and user update habits. The industry has tried to solve this for years with config-driven UI, CMSs, and web views. But these approaches all have a limitation, they require you to be a fortune teller and predict in advance which part of your app might break.  You’re essentially making a bet on which code you think will need to be changed at some point in the future. 

We faced this exact issue at Google while building partial-update solutions for Flutter. These solutions allowed updates to pre-designated parts of the app, but they were limited by the initial prediction of which areas would need changes. This approach often falls short in real-world scenarios.

Believing that you’ll only ever need to change certain code is a fundamentally flawed bet and one of the reasons that Shorebird exists. Nobody, not even the 10x engineer can predict every possible way an application may break or when a cloud outage will occur. Mobile developers need a more flexible and dynamic way to handle updates, one that doesn't require them to predict the future. This is where over-the-air (OTA) updates come into play.

Our design principle is rooted in the belief that developers should be able to address any issue, anywhere in their app, whenever they want.

• Update any line of Dart code wherever trouble pops ups. 
• Add or remove a package.
• Change back-end services.
• Fix dependency issues.

This flexibility is a necessity when outages occur. Outages and bugs don’t ask for permission, you shouldn’t have to ask for permission to resolve them. Shorebird is designed so you don’t need to be clairvoyant and predict where a problem may pop-up, all you have to do is push a patch just like our customer described. Being prepared is better than trying to predict the future. 

‍