At Sliplane, we run managed Docker hosting. One of our core infrastructure needs is to run isolated, fast, repeatable Docker builds for customer de...
For further actions, you may consider blocking this person and/or reporting abuse
I'm curious about Sliplane.io, but the "Why It Was Not a Fit" section has me a bit confused, especially since the "XX-Large" plan only offers 32 GB of RAM. Also, with just four server locations, I'm wondering how distributed Sliplane's capabilities are. It seems like Fly.io and Sliplane might target different needs, so comparing them directly feels a little wrong.
Thanks for the thoughtful response. You're totally right, Fly.io and Sliplane solve different problems. This post isn’t saying “Fly is bad,” just that it didn’t work for this specific use case: isolated Docker builds with large, persistent volumes and heavy RAM usage per build VM.
To clarify: the 32 GB RAM mention wasn’t about our hosting plans, but about the Firecracker VMs we spin up internally for customer builds. Those are short-lived and run on separate infrastructure from customer-facing services.
And yes, we currently operate out of four regions. Although 90% of our servers are in Germany anyway. We do not want to be like Fly with 40 regions, this is a deliberate decision :)
Does that make it clearer?
Yes! Thanks for the clarifications.
Build stories like this always pull me in - nothing beats fixing your own mess and getting stuff running how you want.
Love the transparency here, especially about the hidden limits - I've run into the same surprise costs and quotas elsewhere. Any plans to open source your orchestrator, or is it too tied to your infra?
Not for now! To be honest you could ship a similar version of that orchestrator in a weekend :D
Interesting. I'm a big fan of Ampere and Golang. ARM-based seems to outperform Intel and AMD by miles. Throw in Golang with concurrent processing, and you have a beast. We've halved the number of servers simply by rewriting some services from PHP into Golang and using ARM.
Sadly still on AMD cpus for the most part :(
Great read, been curious about using firecracker directly for a bit. Can I asi if you are running this in top of AWS for infra? or other VM/infra providers?
AWS isnt great because you need the crazy expensive bare metal instances for Firecracker. We run on Hetzner bare metal! DigitalOcean or Google Cloud would also work :)
nice pun in the title
lol, it's so good.
That’s gutsy - dropping the off-the-shelf stuff and just building your own way. Love seeing people actually own their setup, not just patch it up and hope.
This article seems incredibly biased. How could it not be? You’re a co-founder of a competing company.
Oh of course I am based! That doesn’t mean this isn’t true, you can find very similar stories pretty much anywhere on the internet if you search for “fly reliability” or “fly capacity”
Highly recommend