If you've been in tech long enough, you've seen this pattern play out. A platform you've relied on for years starts feeling stale. You notice there haven't been any significant updates for a while, and the tech community isn't as enthusiastic about the product as it once was. Then one day, you see an indication confirming what you already suspected: that the platform is getting phased out.

On February 6, 2026, Heroku announced that they're transitioning to a "sustaining engineering model." The message is short with an emphasis on making sure current customers understand that their current usage of the platform isn't going to change. However, I'm interpreting this announcement as the beginning of the end for Heroku as we know it.

Personally, I've never heard the phrase "sustaining engineering model" mentioned in real life, but I'm going to assume that's corporate speak for "maintenance mode." That means from here on out, customers should only expect the company to do the bare minimum to keep the lights on.

If you've used Heroku for a while, this probably doesn't come as a complete shock. The signs have been there for years. But there's a difference between suspecting a platform is winding down and having the company behind it confirm that they've stopped investing in its future. That's what this announcement is, even if it's carefully worded to avoid saying it directly.

What the Announcement Actually Says

The post from Heroku's CPO is short and somewhat vague. The key message is that existing paying customers shouldn't worry because everything will continue to work as it always has. The announcement says they're focusing on the shift to "stability, security, reliability, and support," likely because their parent company (Salesforce) is following the current trend of all large enterprises and investing heavily in AI. Heroku as a platform doesn't fit this model.

However, what's missing from the announcement tells more than what's in this brief post. There's no mention of upcoming features or improvements to Heroku, no roadmap of what customers should expect coming up, nothing that would excite existing customers. There's just a promise that they'll keep things running as they are right now.

Unfortunately, while your app runs fine on Heroku today, this announcement doesn't foster trust that it will continue to run fine in the long haul. The issue with this post is that a lack of clear updates about the future is rarely a good sign. I've worked in tech for over 20 years, and I know that when a company stops talking about next steps, that silence usually says more than words. I'm not saying that Heroku is intentionally hiding horrible news, like they know they're going to shut down the service soon. But it's also bad for customers to keep quiet about the future.

When a Platform Stops Moving, You’re Moving Backward

Heroku attempted to soften their announcement by reminding customers that things aren't going to suddenly grind to a halt. You'll still be able to deploy, spin up new dynos, and have your apps working as normal. But for any platform in this situation, just saying that what's already functional will remain functional shouldn't be enough.

While Heroku stays in their "sustained engineering model" phase, all their competitors will keep moving forward. Just think of all the changes that have happened in the world of software development since Salesforce acquired Heroku in 2010. We have a ton of options for cloud computing services, and they've gotten much cheaper over time. Deployment tooling has gotten drastically better, which was one of Heroku's main draws. Both of these areas will continue improving over time, but based on what Heroku is saying, they won't be keeping up with any of it.

I've used Heroku for both personal and professional work, starting way back in 2008 when they were focused on Ruby on Rails only. If you've logged into a Heroku account recently, you already know how the platform has felt stale for a while. The dashboard hasn't meaningfully changed in years, and the removal of the free tier in 2022 was the first loud signal that Salesforce was pulling back. That decision drove away the developer community that made Heroku popular in the first place.

Now, with their current plans of going into maintenance mode, it'll accelerate all of these issues. The gap between Heroku and modern alternatives isn't going to shrink. Instead, it's going to widen. It's been widening for a long time, so just imagine how large the gap will become now that there will probably be a small crew working on the platform.

This Is How Platforms Die

I want to be clear: I'm not saying that Heroku is shutting down tomorrow and you should abandon ship immediately. The announcement explicitly says otherwise, and I believe them. Your apps aren't in imminent danger, and you don't need to flee as soon as possible. For some teams, staying on Heroku will be worth it to them.

Still, going into "maintenance mode" is a pattern you've probably seen at your company or in products you use. Someone decides a product isn't worth further investment, so they frame it as focusing on stability, reassuring you that nothing will change. And then, slowly, things will degrade. One day a feature in the app stops working because it's not supported on newer browsers or operating systems. Then, when you reach out to customer support, they don't respond for days. That promise of focusing on stability suddenly becomes a broken one.

Often, we remain on a specific app or platform even when there are better alternatives because the cost to switch is high. But when those services become neglected, suddenly the cost of staying begins to outweigh the cost of leaving.

Alternatives become so much better that the cost of staying starts to outweigh the cost of leaving. Additionally, because it's a slow decline, it's easy to push off migrating off until one day you realize you're stuck on a platform that's fallen years behind and the migration has only gotten harder. The worst time to migrate off any platform is when you're forced to. Rushed migrations lead to cutting corners, and cutting corners on infrastructure is how you end up with production outages and angry customers.

What You Should Do When Your Platform Shows These Signs

If you're on a platform showing the same warning signs as Heroku, I don't think you need to panic, but I do think you need a plan. Here's what I'd recommend:

Figure out what you're actually paying for

Pull up your monthly bill for the service and look at every line item. Many teams are paying for add-ons they barely use or could easily replace with self-hosted alternatives. Understanding your current costs gives you a baseline for comparing alternatives.

Pick a realistic timeline

Don't try to migrate everything next week, but don't put it on the "someday, maybe" list with no specific date either. Depending on your application's complexity, a 2-4 month timeline allows you enough room to do it properly without letting it drag on indefinitely.

Test alternatives before committing

With plenty of viable options out there, it's essential to do a little legwork to discover which provider suits your situation best. For example, you might want to find what the best price-to-performance ratio is for a typical web application, as I did in this article. Investigating the alternatives will help you find what you need.

Look at modern deployment tools

One of the reasons people flocked to services like Heroku is ease of use. For instance, having your app deployed by running  git push heroku main is hard to beat for simplicity. But nowadays we have alternatives that are just as straightforward. For instance, tools like Kamal make deployments just as easy as Heroku, with additional benefits like using your preferred services. The developer experience gap between older and new platforms is much smaller than it used to be.

If you don't have the bandwidth, get help

Infrastructure migrations aren't a great place to learn by trial and error. If your team doesn't have experience with this kind of work, bringing in someone who's done it before can save you weeks of troubleshooting and prevent the kind of mistakes that are expensive to fix in production. This type of work is what I do, but whether you hire me or someone else, don't let a lack of in-house experience be the reason you stay on a declining platform.

Wrap-Up

Heroku's announcement isn't a "drop-everything-now" type of emergency. If you're using the platform to run your web apps and are happy with how things are, there's no issue remaining where you are. However, the announcement that Heroku published is a clear signal about where the platform is headed, and it seems like it's headed nowhere soon.

If, like me, you believe that Heroku as a platform is going to fall behind and stagnate, it's a good time to begin exploring the alternatives. The best time to leave a platform is when you can do it on your terms, without pressure and with enough time to test everything properly and get it right. The longer you wait, the harder the move becomes, so start planning your move.