Posted: 2026-02-24
On the SporeStack side of things, I resell DigitalOcean VPSs. They work great and they're very reliable.
Anyway, I noticed this interesting headline on DigitalOcean's blog.
I'd say the post is well-written and an effective sales piece for people on Heroku right now. But I also feel like it shares some deeper truths about what most people in the tech industry think.
Heroku's move to a "sustaining engineering" model was carefully worded. It avoids the term end-of-life. It reassures existing customers that nothing changes immediately. It emphasizes stability and support.
But if you read between the lines, the message is clear: there is no forward roadmap. There is no new feature investment. The platform is entering maintenance mode.
For teams running production systems, the real risk isn't Heroku going offline. It's standing still while the rest of the industry moves forward. Platforms rarely become risky overnight. They become risky when innovation stops. Teams need runtimes that stay current. They need security patches that ship on time. They need platform support for the frameworks and the tooling they're actually adopting. When a platform is no longer investing, those gaps widen over time.
I am certainly not against all innovation. There are many areas where forward progress can be made. However, this assumption that innovation is always mandatory, that nothing is ever "good enough," is concerning. There is the phrase about "needing security patches that ship on time" -- not developing on features obviously doesn't have to mean that security patches don't ship quickly.
In fact, features often create new security holes. Often, old code that is battle tested and patched is most efficient.
Same goes for technological innovation. I think it's amazing to be able to converse around the world and share ideas. We've been able to do that for many years at this point over the Internet. How much more "progress" do we need?
In dabbling with Perl, I realized that I had no memory of a major Perl vulnerability. It's not to say that they aren't and that there haven't been. And certainly adoption is a big part of it. But I bet that it's largely because it's well tested and matured. By no means the latest and greatest, but can that not mean that something is "good enough?"
Now I'm not implying that everyone should use Perl. I can understand not using it. But I think it's an unhealthy mindset to think that things are risky as soon as they stop innovating. I do think there are some who just "check out" and stop trying in general. I don't advocate for that. What I don't support is the notion that everything must always evolve. It's like everyone is racing against eachother. It's not just a healthy capitalism anymore, it's an obsession to be able to make everything we once did obsolete, and "innovate" faster than ever before. It's now even to the extent of wanting to optimize humans out of the pictures altogether. Why on earth would an apex predator secede his position? Let alone intentionally make his successor! (I'm describing the massive push for AI superintelligence.)
Though I'm no particular fan of Heroku, in some ways going into "maintenance mode" means that you've succeeded. You made something that worked so well for people that even after the hype is gone, you're there to keep it going. People keeping using it because it works, not because it doesn't.
Again, I'm not against all forward progress. I think it isn't good that Slow Servers relies on one router that *could* fail at any moment. Is it likely to fail? No. But basic redundancy, that we've had for... 20? 30? years isn't a crazy thing. Especially when it means I have to drive an hour an a half (or more?) to mess with some gear, maybe in the middle of the night. But this is really just getting it to a standad that we have known for many years and hasn't really changed.
Thank you for reading.