An Update on Heroku

> They took all of the great things about Heroku but also dramatically improved and added new capabilities

I also love Fly, but they were missing easy managed databases (which always seemed like the main reason to use Heroku to me). And now they have them they're very expensive (even compared to Heroku). Which is a shame because their compute is very cheap.

I think the "Heroku story" was less about technical limitations, but everything except technical limitations. More than a decade ago, I started learning and building on Heroku and hosted all my side projects and client projects on Heroku. Then when they got acquired, I was naive; then they removed their free tier and that broke my trust.

I primarily worked on PoC/MVP development where I worked to bring ideas to something barely tangible. And Heroku's free tier decisions meant it was a barrier for developers to develop on their platform. Pay first, develop later. It was like the rest of the industry.

After that, I just exited containerized platform-based application development entirely because convenience and having that weird developer philosophy "I must not pay because I can find a way" was less of a reason than sustainability. For me, containerized application platforms was about POC and MVP. If there was growth then me or the client can pay for the convenience. But if there was nothing, pretty easy to delete the project.

Then I committed to replicating the Heroku experience with a small VPS, backing up via rsync, and moving from PostgreSQL to SQLite. I can even charge clients for hosting (+ maintenance) on my VPS.

I do not know, to me containerized application platforms are limited by commercial challenges rather than technical ones. I see tons of containerised application platforms, but the trust has eroded because of a single company.

I have changed my development facility and laid the groundwork to not commit to these platforms. Sustainability over convenience.

Sure, I understand and respect folks at fly.io, render, railway, and even the open source variants of these companies (Caddy etc.). But there is no sustainability guarantee for these platforms. It was not just about the "free tier", to me it transcends to a philosophical point about building applications in general. Sure, there could be a new era with AI making MVP/PoC development easy through hosting in containerised applications, but that is a tangent point.

If Heroku were doing everything right, there would not be a dozen application platforms out there, but they made mistakes and, in my opinion, made the entire containerised application platform model untrustworthy.

I have moved all Rails apps away from Heroku in the last years. It was great 10 years ago, but then became expensive, full of bugs and with terrible support. All our Rails apps (Pushpad, Newsletter.page, etc) are running on Cuber gem + DigitalOcean Kubernetes... In the last years we achieved 100% uptime (five nines), zero subtle bugs and huge cost savings.

This is surprising to me. I actively used Heroku during the early to mid 2010s. I do not remember ever seeing the Salesforce logo there much later.

For what it's worth, I've been very happy using Crunchydata PostgreSQL with Fly.