Still working on upgrading our server infrastructure so the app can handle millions of users. We’ve had to solve two issues over the past couple of months: 1) server capacity and 2) app speeds.
1) Server capacity has been a real challenge as we’ve had to reimplement most of our features from the ground up. Every day we’ve worked to remove our old architecture and implement new architecture that can SCALE. When we started JellyChip 3 years ago we never expected implementing a feature like Lifeline that requires such responsiveness; crowdsourced giving and the need for instant feedback. The same applies to a fully-fledged Chat feature and Snaps.
We probably also didn’t anticipate the huge number of users that could potentially use JellyChip over the next couple of months. This has been a big thing for us and we’ve had to prepare for that (otherwise the app would die).
2) App speeds. The real issue for us has been removing 3 years of old code. Imagine driving your car for 3 years without a wash. Or wearing the same clothes for 3 years without a wash. Everything deteriorates, becomes disorganised, smells and needs a clean. Even our code smells sometimes.
In fact it stinks. We’ve spent the last 8 weeks cleaning our code (refactoring) and that’s meant untangling the mess of 3 years of features that were not implemented, differentiating different developers, updating our documents and deleting old data. Coming towards the end of this has been extremely satisfying (particularly the last week) which means we are now much more well-placed to build new features. A cleaner code foundation means a much shorter time to implement new features.
So we’ve come to the end of 1). Now server capacity is an endless journey but we think we can be prepared now for 500,000 users in our database and 3-4,000 people active per gift. This is pretty awesome.
Over the next couple of months we’ll be able to focus exclusively on just bringing more AWS instances online to grow our database and active user count.
I mean, if JellyChip goes off the clock and gets millions of users at once - then we’ll have to have a think.