At one point I used AWS for commercial work and Glitch for fun and experimentation.
But AWS makes everything [except bucket-based storage] a giant royal pain in the butt, with a million layers of abstraction and documentation that is written for those who already know what it is explaining.
So I found myself publishing some serious stuff on Glitch. It works now, I figured, and I’ll worry later about scaling. [Later is here, brothers and sisters!]
But I have one large-scale project that needed a complete re-architecture. Way too big for Glitch, so I tried moving ii from AWS to Google’s platform (Firebase).
I like. Google Cloud Storage is as easy to use as AWS and features like IAM are a million times simpler, especially if you want to collaborate.
Simple static hosting on Firebase, is super easy and pretty much free. Unlike Glitch, I can stay in VSCode and instantly deploy to Firebase hosting and push to GitHub repo, while managing my assets very directly. Frankly it is so easy that, without thinking about it, I stopped using Glitch for quicky hosting. Firebase is free, easier and more secure (repo-wise) and so much better performance. (Glitch, I still love you.)
Also you get user authentication for free (I use it) and tons of tracking features that I do not use as well as a million infrastructure features I don’t even understand.
It gets hairy when you want to move your Glitch Node apps.
I have ported small Glitch Node Express servers to serverless Cloud Run functions (These run Express under the hood, it turns out) – and this is relatively easy. With enough work. In fact I usually edit these in the cloud, like with Glitch (Cloud Run functions do not play nice with VSCode and Github😒) .
Serverless functions may be impractical if you have a complex server. I’m about to find out.
But the huge problem with Firebase is no SQL support. I ported my big app from SQL to their NoSQL database but it was a huge job with a lot of rethinking. But the result is worthwhile. The new database is better in most ways.
I have an active public app on Glitch now that uses the wonderful SQLLite. It is a complex app with many daily users who submit and discuss some important UGC. Database is the core of the app and I really do not want to rebuild it. I don’t think I can in this six week window.
Any ideas where to go with a system based on SQLLite3??