Heroku login down for 7 hours
This, I think, is finally the last straw for me.
I've already moved my databases off heroku long ago, tried many various PaaS (render, fly) and tools (kamal) for VMs on places like DO and Hetzner. All my newer apps are running this way with lots of learning curves.
I just can't fathom an enterprise business being down this long.
Rest in peace, Heroku. You were great once. You can get Heroku/Vercel/Render like experience on your AWS account with rock solid support from AWS. Checkout https://localops.co if you guys are migrating away from Heroku.
(I'm the founder of LocalOps here) App runtime: Connect your Github repo and AWS account in 1-click and get Heroku like-PaaS automation. Just push to your Github repo and we will deploy your code automatically. Zero cloud config or DevOps effort required. Databases: Just add the databases you need on a json file and they will get auto-provisioned. AWS managed RDS (Postgres/MySQL), Elastic cache (Redis/Memcache). No need for manual provisioning or maintenance. Zero downtime deployments. No ceiling: Scale infinitely on your AWS account Cost: 60-70% overall cheaper than any traditional PaaS alternatives like Heroku/Render because you are paying for infrastructure provider directly for all servers and paying LocalOps just for automation. Support: Get rock solid support from AWS for infrastructure uptime. It's hard to beat the stability of AWS. I've not once had an issue getting into their console or managing my resources. I've been using them for 15 years at this point. They're the only big cloud provider I feel comfortable with. Route53, EC2, S3 and WorkMail are some of the best products available in the market right now. If you value your time, the additional premium is almost certainly worth it. AWS is the closest I've ever seen to "set & forget". “They're the only big cloud provider I feel comfortable with” Firstly do you work for aws, as saying the other big cloud providers are inferior is unusual. Azure for example powers a lot of critical enterprise which are Microsoft.net centered. And secondly there are lots of other providers outside of the big cloud providers who can maintain 99.99% service levels. That's a big one and I'm sure will be the final nail in the coffin for plenty of people. No reason to use Heroku in 2025, especially for Rails apps (you mentioned Kamal). I moved my app onto Hetzner + Cloudflare R2 and I couldn't be happier! https://olly.pagecord.com/migrating-the-pagecord-rails-app-f... I recommend Cloud66 over Hatchbox which is more production-grade, especially for DBs. Yeah I'm operating on a shoestring but tbh servers are super-reliable as is Postgres. What even is "production grade"? Read replicas, point-in-time restores etc? For a product that is ~$1000 ARR I think I can live with a bit of risk (I do regular backups of course and have a restore process). Maybe it will bite me in the ass, but I'll risk it :) Happily using Cloud66 but a strong recommendation for Northflank here. I already had a running cluster there and migrating a very complex Heroku app over only took 20 minutes today. (Would have been 5 if I hadn't screwed up .gitignore...) I'm currently migrating to fly.io, your thoughts? I want to avoid K8s for the initial stage.