Is Heroku dead?

With Render & DigitalOcean App Platform, I think Heroku is pretty much done these days. They still might have use cases for the enterprise though.

With a wealth of integrated tools and a simpler (although more costly) transition route to business features like HIPAA and PCI compliance, Heroku's attractiveness lies in its ability to allow us to concentrate more on customers and less on infrastructure. It's a tried-and-true platform that needs little effort and expertise from the developer. It's natural to think about more recent or less expensive options, but since our clients have so many other preferences, it's difficult to defend infrastructure optimization.
mapquest driving directions

The beauty of Heroku is that it lets us focus more on customers and less on infrastructure, with lots of integrated tooling, and an easier (but expensive) migration path to enterprise stuff like HIPAA / PCI compliance. It's a proven platform that requires minimal devops time and skill by the developer. It's tempting to consider newer or cheaper alternatives, but it's hard to justify optimizing infrastructure when there are so many other things our customers want.

I try to deploy my four differrent sample nodejs on Heroku 8 times. All are deploy successful but API not found everytime. Success rate 0%.
Then i need to purchase digital ocean vps plan and deploy and run same code , it works.

No, i do not think so. Notwithstanding the alternatives mentioned below, there is NO SINGLE service that does what Heroku does excellently. If you need to replicate Heroku's functionality you end up glueing lots of other stuff together.

I don't think it was ever really viable, it may have been the best thing out there 10 years ago. But since AWS came around and Salesforce stepped in the way, you can be sure it is a done deal.

Heroku is dead.

If you need to get something out there quickly, a hobby dyno might still be useful. bitlife is a life simulation game in which players assume the role of a character, make decisions on behalf of that character, and then observe how those decisions play out over the course of a simulated lifetime.

Hi, Yaron here (-:

Co-Founder and Chief Architect at yuvital.com
We have been happy customers of Heroku for around 5 years but have started moving to Render.

The first big difference is the price. We have enterprise customers with medical sensitive data and we needed to use Heroku Private Space. This is extremely pricy since you pay a base price of $1000 just to get stable outbound IPs. In Render, you get this for free.

Also, when you are in a Private Space the cheapest dyno is around $70. This is very limiting because sometimes you just wanna a small dyno to do some random work or be up.

Another nice feature of Render is infra-as-code. You change a YAML file, and the infra is updated.

Also, env groups are great. You define the env group once and attach it to how many services you want.

Huge win - "real" auto-scale by memory and CPU. Heroku scale is based on latency. But most of the time, it's your database that is choking and causing the routes to have long latency. This has nothing to do with the dyno's CPU or Memory.

In Render, you finally get an auto-scale that monitors the server's actual memory and CPU (avg across all instances).

Also, in Render you use GitHub as a source and not Heroku's internal git. Github is much more stable. Heroku's git if often down or very slow to push.

Generally, render feels more snappy in builds and deploy times.

And of course, it's active and growing.

Heroku still has the upper hand in features and community (marketplace) but Render will catch up, and for now, all the things listed above are already enough for us to make the move.

Happy to hear your thoughts.

Hobby dyno might still be good for something experimental if you need something out there quick.

I'm always surprised by the comments to this question. We looked all around for other hosting options that let us focus the most of our time building, and the least on devops. While there were some other competitors (Fly.io and Render being the closest), we haven't seen any notable reason to move to them over Heroku other than price.

In our case, Heroku does what we need. It's battle-tested and works. We don't spend much time thinking about devops, hosting, or setup. Eventually we will, but that's less important than product validation at this point. If you're able to do that on a cheaper platform that takes less time for you to use, then go for it.

we started using vercel instead. much better experience!

Heroku is not viable for a very long time already.

That's why you should pick platform that works best for your stack of choice.

We've built https://appliku.com/ for Python/Django

There are others too: https://twitter.com/kylefox/status/1530256380744978432?s=20&t=z6xvsTAiXnUDrqxEBTMK9w

Heroku was a really great alternative and still is, but although the Heroku platform is easy to use, it's built on top of Salesforce, an incredibly outdated and legacy system.

Salesforce isn't going to change that ever, as they only care about Enterprise clients that don't have another option. Because of that Salesforce will prioritise the enterprise clients and will continue to neglect users like us. And fixes new features will be few and far apart.

Thanks for everyone posting alternatives, I've also been looking to switch but haven't found anything that's as familiar as Heroku.

I recently moved to the railway.app from Heroku. I would recommend it since its servers never restart or idle even on the free plan, unlike Render or Heroku.

We're building an alternative for it and launching it soon. If anyone is interested to get access to our Alpha version let me know.

Maybe not. They had a cyber attack related to GitHub integration. In their blog they had clearly stated that they will not resume their service unless they've fully fixed the issue.

It's definitely dying slowly. This is how most of companies end up when they are acquired by Big companies.

It's not dead, but it seems to be just hibernating.
For example, their 4 years old article states that their ACM letsencrypt solution does not support wildcard subdomain certificates. Letsencrypt added support for this 4 years ago. It's a relatively small change to make, and they are unable/unwilling to do it.

I love Heroku, and I don't think there is any other solution at their level yet. But I'm sad to see it went on hibernation mode more or less after being acquired by Salesforce.

P.S. What are good Heroku alternatives nowadays?

They have not created anything else or added new services, they stayed in time

What's going on at Heroku? They seemed to be thriving and suddenly all this.

With the way things are going, it seems the answer is increasingly "yes".