Hacker News new | ask | show | jobs
by justjake 1512 days ago
Warning: There are a couple people in this thread mentioning products that they're investors in. I won't call them out expressly, but I think people should be direct about their incentive alignments. It's disingenuous and borders on astroturfing IMO.

Background: Founder of Railway.app here.

There's a lot of these companies popping up that offer a "Heroku replacement", and once you dive in, you realize you have to pay $300/mo for a Kubecluster + $200/mo for the wrapped Kube service

In our experience, people move off Heroku for a couple things:

- Cost: Kinda self explanatory but Heroku pricing ramps hard

- Flexibility: Heroku's not great for anything beyond stateful monoliths

- Scalability: Notoriously Heroku's SLAs aren't that great

In my mind, you don't replace Heroku with a minimum $500/mo Kubernetes cluster. Not only is this cost prohibitively expensive, but Kubernetes itself is a jet engine, and if you're not trained to use it correctly, you can risk catastrophic failure (on costs ballooning, on dataloss, etc)

We're working hard at Railway to provide not just a Heroku replacement, but a next generation, composable infrastructure canvas. $5/mo, 30 seconds, and you're up and running.

Demo:https://user-images.githubusercontent.com/5499880/165187948-...

Would love any feedback and thoughts people have about Heroku, my thoughts above, and what we're building :)

EDIT: Wow, just went from 11 -> 8 upvotes. I suppose they didn't like the astroturfing callout

EDIT2: Render isn't who I'm talking about. They're great.

8 comments

Personal criticism of Railway: every time I emailed Railway to request to be unsubbed from the newsletter and my account fully deleted without a trace no one ever replied (the unsubscribe link in the newsletter was broken). It left a really sour taste in my mouth.
Hey Emptysongglass,

Angelo here, Support Engineer from Railway. I am sorry to see you go from Railway and personally speaking, I wouldn't want my PII floating anywhere I wouldn't want it to be.

If you can email me directly: angelo (at) railway (dot) app with your username from HN with your account email. I would be more than happy to make things right here.

Sigh. I swear I emailed everyone who ended up in that state, but it might have gone to spam cause well, we had this issue indeed.

We had some pretty massive issues with an email provider. Have since switched to Postmark. Apologies there and totally understand that absolutely blows.

They’ve been shilling their R•nder and P•rter services on every single thread about this Heroku incident, and it’s frankly a problem that HN moderators should address.
It's not just them sadly. I think the only one who I've seen who isn't guilty of this is Fly.io.

They build cool shit, they blog about their cool shit, because it's cool shit.

(Render founder) I haven't seen a single Render employee or investor post on HN about the Heroku incident; this comment is the first we've participated in any discussion on the topic. I'd love to see links if you can share.
Well here is you (with your disclosure) participating in a discussion on the topic a few days ago:

https://news.ycombinator.com/item?id=31053504

You’re as capable as I am of searching HN for mentions of your service.

He's responding to someone who specifically mentioned their experience using Render. I don't consider that shilling. I personally like seeing that kind of direct engagement from founders on HN.
Yeah, that’s not shilling, but it is in direct opposition to what they stated in their comment here about not participating in any discussions about Heroku.

I’m calling attention to my inability to trust this person and their company’s marketing tactics.

What should HN do? People are free to downvote it until it's dead. That seems a more democratic way to handle it.
I'm such a HN normie does HN even have a downvote button? I can't find it
You need 501 karma points to downvote!

Here's a list of all hidden HN features: https://github.com/minimaxir/hacker-news-undocumented#downvo...

Oh wow I had no idea thank you!
Your pricing page is confusing and I can't find enough details about the offerings.

1. What happens if i exceed the outbound bandwidth quota? 2. In Team, what is "seat" ? Also pricing number of team members is a bit of a turn off. I don't expect that kind of pricing from an infrastructure company. 3. What if I need more than 100GB disk in developer plan? 4. The postgresql offering - Does it offer failover, replicas, PITR etc? 5. How are logs handled?

Railway looks really cool but what is the procedure if you want more than one of the same data store in my case one redis for cache and a second redis for aggregating usage metrics?

Edit: I found a small bug, I can't link my discord and railway.app accounts because my github is < 180 days old, it says that I must enter billing information but I did that a couple hours ago. There's no way to post this message in discord since private messages to your team members get rejected and public messages require that linkage.

Hey Beninsydney,

Angelo here (Support Engineer) from Railway. I was personally responsible for implementing the whole flow to link your account to post a support question. I apologize, I put those limits in place to help fight a recent wave of spam. If you can email me directly: angelo (at) railway (dot) app - I would love to answer any questions you may have.

In the meantime, I will be sure to make things right for your account. For others who might face the same issue as you, I will rectify the issue with the support flow moving forward.

I agree 100% on the misdirection happening on many of these so-called Heroku replacements.

Offering me 'kube-as-a-service' over a layer of AWS or whatever on top of some massive cloud provider isn't really a heroku competitor. Thats just throwing some plywood over a ball of mud and trying to offer some hand holding with all the rough edges. I don't wanna have to decide between AWS or GCP (see one of the first docs on Porter: https://docs.porter.run/getting-started/provisioning-infrast...), or even think about them. I don't want to learn k8s to spin up some side-projects, and ultimately if you go with one of these the abstractions _will_ leak. I don't wanna go look at AWS UIs to see the status of my database (see https://docs.porter.run/deploying-addons/postgresql#persiste...).

The fact is if you need k8s, then you probably need a team of smart people setting it up and continually managing it who are very knowledge about your app and its specific needs, because you better have the scale that requires. Because if you are spinning up a k8s cluster for a monolith or a few microservices w/ 200 customers, you are just going to burn thru dev time and cash.

Heroku has been constantly praised since it took off because it does massive amounts of things behind the scenes to abstract away all the operations you don't want to worry about. You can launch a tiny prototype or small to medium startup in under an hour. You can add a DB and Redis w/ snapshots and automated backup and monitoring all w/i the Heroku API or UI. Does it expose the full power of the datastore and let you do everything you can do with RDS or a VPS? Of course not, but thats a totally valid trade off when you are just trying to get something shipped to see if it has traction.

So if you are looking for Heroku alternatives: know that things like Porter aren't really direct competitors, no matter how much they are marketing themselves as. From what I've seen Render and Railway are much more in line to be the next Heroku-replacement.

edit to mention, since the parent brought it up: I have no horse in this race. I'm still a fan of Heroku and use it daily, but don't work for anything to do with dev tools or PaaS.

Would you be willing to go over our Heroku environment to see if we'd be a good fit for Railway? I can't tell from the demo if you have things like automated backups, rollbacks, workers, add-ons for logging/monitoring/performance/etc. Those are all tablestakes for us. But Railway certainly looks great.
I would love to!

I can tell you a couple things off the bat:

- We have automated backups but they're for our own internal disaster recovery. User backups are something we want to do but haven't put on our roadmap yet

- A key thing with Railway is "It's just code". So, a worker is just another service. We don't have special casing for specific types of code. We just run the code! So, yes we do support them

- We have a lot of stuff built in, but we also support deploying say, a containerized DataDog agent, or a Dockerized sidecar, or application level Sentry integrations

We have a very particular vision of how this stuff should be done, so it's going to take some time. My promise is to always be super honest about the platform, so it would be more of an "Customer Discovery" vs a Sales call

If that all sounds well and good, you can email me at jake@railway.app! Offer goes for anybody interested but again, I'm not here to push the platform just to gain clarity on what we're building :)

I was trying out your free tier and was looking for any way to run shell commands (like heroku run). There's "railway connect" to connect directly to a database but really what is needed (i.e. table stakes) is to be able to do something like "railway run ./manage.py shell" remotely (NOT local).
Just saw someone else linking to railway.app, curious how does the scaling work?

The pricing showed a lightweight website priced at less than 1 CPU / month -- is there some undocumented fractional scaling going on?

I was looking at your site earlier due to link up the page, and I have a question about the pricing:

What exactly does 0.1 of a vCPU equate to?