For about 18 months beginning in spring 2015, I ran a newsreader web-app by the name of FeedHop. It parsed it's last news item in November of 2016.
I built it because I wanted to learn something new (in this case React and Go), and all told it went better than I had planned it to. Pretty soon the app was parsing news items from thousands of feeds. Despite my best efforts to rewrite code and generally make things more efficient, server costs continually increased and I had no way of covering them.
The months rolled by, and it slowly became clear there was only one thing to do. FeedHop had to be laid to rest.
Services are inherently things people rely on, so there's an expectation they will always be running. But of course to keep a service running requires time and money. Good will can only go so far. Not accounting for this was one of the biggest errors of judgement when I started out.
I believe if I had open sourced FeedHop from the start, it could have grown into a more sustainable project. It could have been self hosted, and it could have evolved into something new. A few people suggested I make the app open source whn it closed down, but over time the code had become such a tangle of tweaks and hacks that it was (and is) a long way from being fit for public consumption. Open sourcing at the start of the project would have avoided this.
The sad fact was that FeedHop was something that I and others enjoyed using, and that I couldn't find a way to make it work saddens me. It would have been amazing to see it grow, even if it wasn't in it's original form. I still hold out hope that maybe one day it will.