28 comments

[ 2.6 ms ] story [ 69.2 ms ] thread
How is it different than Dokku?
Perhaps to answer the question, one might ask: How is Heroku different from Dokku?
For one, Heroku has dedicated customer support when you pay them.
Right, dedicated support should be available at launch. Thanks for the feedback.
That difference is very clear: you still need to buy and manage your own (cloud) server to run dokku on.
Having used both extensively, the answer is as simple as one is private/hosted and the other is open/self-hosting.

Heroku is easy, but expensive. Dokku is more work/management, but cost-effective.

Looking forward to seeing this mature. I am using dokku to host Webase [1] on a dedicated server and also have used Heroku extensively. It feels like there is room in the market for a tool like this.

[1] https://www.webase.com

How does this differ from CloudFoundry?
(comment deleted)
See also: CapRover. Beats dokku all to hell, in my experience.

https://github.com/caprover/caprover

Be careful, their (CapRover) embedded NetData image has the NetData upstream defaults for spyware, so you are advised not to enable it (use of their embedded NetData is optional).

https://github.com/caprover/caprover/issues/553

It is straightforward to set up CapRover with an ssh deploy key and a webhook for deploying automatically on any repo change (even from a self-hosted Gitea, which CapRover also makes easy).

The github issue you linked to and the discussion it contains is the reason I am sticking with dokku. The CapRover maintainer seems to argue in bad faith and not have the users best interest at heart.
Agree. It does not inspire confidence to recommend a product and quickly follow it up with worries about spyware.

And dokku works well for my use cases.

Author here. Just to be clear, this is false information. It's a toxic thread.

1) The author of that thread is calling a software with analytics a spyware. By that definition, any website you use, including this very website you left a comment on, is spyware.

2) NetData is built-in, but it's not enabled by default. Use is free to install/ or not to install the package.

Silent surveillance, which is not disclosed to the user. Additionally, the user's IP (and thus location) is also provided to Google with the surveillance data.

Webpages and local software are not the same thing. Local software does not need to report my activity to Google, a third party, to work. Webpages do require second party communication to function.

Surely you realize the difference. Your attempt to conflate the two is not productive.

If you believe they're the same thing, let's call it "analytics" then, exactly what it is. Let's avoid using the terms like spyware or etc which is an umbrella term which covers very serious damages such as stealing your bank info.
Transmitting a user's activity silently and without their consent is called spying, not analytics.

The way that such surveillance is legitimized into not-spyware is via obtaining AFFIRMATIVE consent from the user.

Without that, it's spying, and software that does so is spyware. It's quite simple.

Isn't Show HN for products that we can actually try out? Looks like this is just collecting emails at this point.
What part of this is open source? I don't see any Github or other links to said source.

It looks like you're going to eventually build a company around this and maybe you will think about releasing the source then? In which case, the title is still misleading.

I do appreciate that you've released documentation ahead of the product, it does indicate that you understand how important that aspect of the business will be.

It's great to see all the good competition in this area. The more use, discussion, and cross pollination among them the better. Here's another one (Akkeris) I've used, open sourced by an excellent former employer and devops team: https://beta.akkeris.io/
What's the point running something like heroku on your own server. The whole reason to use Heroku is that they take care of a lot of stuff you don't need to like updating servers, certain aspects of security, making sure there is no downtime and so on. If you run it on your own server you need to take care of all those things yourself, then why not use just a simple aws instance?
Neat, although apparently not available yet.

In the meantime, if you want something that is simple enough that you can read the entire code base in one sitting, check out Piku:

https://github.com/piku

People at Convox[1] and Remind101 are doing similar open-source work.

Convox's business model is selling you the managed version of their open source rack. We've been using Convox in production for over 3 years now and we've had pretty good success with it. It's not perfect but the UX improvement for developers when building docker images, promoting new builds or rolling back cannot be understated.

Remind101 has built Empire[2] which is not a product they actually sell but rather the tool they use internally and have decided to open-source it. Empire is a little less managed and while it allows you to run services on it it requires you to have your own solution for building and hosting docker images.

[1] https://convox.com/, https://github.com/convox/rack

[2] https://github.com/remind101/empire

I might be wrong, but isn't Convox only operational in the AWS cloud. If I want to run Convox on Azure, GCP or bluemix, will this be possible?

The empire alternative pointed out also appears to be another option only operational in the AWS cloud.

Correct, both the Convox ECS rack and Empire are AWS exclusive. Convox has a public-beta of their rack running on Kubernetes which is cloud agnostic as it only needs a Kubernetes cluster to operate in.
Does Convox provision and manage the underlying Kubernetes cluster or depends on the Cloud provider's managed Kubernetes offering?
Convox does not manage the underlying Kubernetes cluster. If you're using EKS it's managed by AWS, GKE by GCP, AKS by Azure, etc. If you're using a self-hosted cluster, then it's up to you to manage it. The Convox platform runs as Kubernetes CRDS so it is cloud agnostic.