rack VS Flynn

I would encourage anyone looking at porter that needs more flexibility, a deeper feature set, a more extensive cli, and/or flatter pricing structure to check out convox [https://convox.com]. I've been using it for years on top of EKS and have been able to defer hiring a dedicated dev ops person.

n8n is a  powerful workflow automation tool that every business can use to work more efficiently. However, self-hosting this solution in an efficient infrastructure platform like Convox is the only way to explore the full potential of n8n’s capabilities. With this tutorial, you should have a step-by-step practical guide showing you how to do this.

In that case, there’s a need for more efficient methods for hosting PHP apps from the local machine directly to the production environment. Using a modern and user-friendly Platform as a Service (PaaS) solution like Convox helps you easily deploy and manage PHP applications on all infrastructures, including AWS, Google Cloud, Azure, and DigitalOcean.

Convox is one of the best PaaS platforms for growing businesses that do not yet have the resources and capacity to handle a large team of DevOps engineers. The platform offers a perfect solution for companies that manage multiple client projects with a limited number of DevOps engineers onboard.

I don't really have the resources to re-architect my application to use Lambda, Fargate, etc. so I just want to find some quick wins within the current constraints. I use Convox.com, which uses Terraform to spin up and manage the EKS cluster. (They provide a CLI and web UI similar to Heroku on top of AWS EKS.) It works really well most of the time and makes it really easy to achieve HIPAA compliance, etc. So I'm pretty happy with it, but I need help understanding it and optimizing it. We can easily tweak the underlying Terraform config to make any changes to the infrastructure.

If you’ve read this far but actually haven’t managed to get started with Kubernetes, this is the toughest part of the process - but don’t worry, we have you covered. The good news is that Convox (a PaaS solution for Kubernetes-based deployments) can help you get started with Kubernetes in just a few clicks.

Convox for example offers your organization direct Kubernetes access and removes the complexity of managing and operating your Kubernetes cluster directly. Furthermore, Convox is a multi-cloud product that is compatible with several cloud service providers including DigitalOcean, AWS, Google Cloud, and Microsoft Azure. It is free to start simplifying your application development process with Convox and it only takes a few clicks to get started.

This case study lifts the hood on one such development team that has had to adapt its PaaS choices with time. We will discuss the teams’ journey towards automating its infrastructure and deployment process through a series of migrations - from Digital Ocean to Heroku, to Convox. This rare insider’s perspective mirrors the processes, problems and solutions faced by similar teams along the journey, drawing lessons and teachable moments for software development team managers.

Instead of manually configuring Kubernetes yourself for app development and deployment, you can use Convox. The Convox platform is designed to simplify your infrastructure management tasks. With Convox, you and your team can smoothly deploy, scale, and manage applications effortlessly, exploiting the power of Kubernetes, without the headaches.

Considering other orchestration tools like dokku, dcos, deis, flynn, docker swarm, etc.. Kubernetes is no where near to them in terms of lines of code, on an average those tools are around 100k-200k lines of code.

> we are indeed writing a new orchestration system in Go, called `flyd`.

I know it's just a wild coincidence, but I couldn't help but be reminded of https://github.com/flynn/flynn

I love these projects and I'll definitely look into it more later. At the same time, I think it would be hard for me to want to leave Kubernetes. K8s isn't something that I love, but it's something that I know works and will continue to be supported in some fashion. I've seen Flynn (https://github.com/flynn/flynn) EOL. I've seen Docker Swarm's future be a bit hazy. I've seen Porter decide to put pretty restrictive limits on their free tier (though it is open source). I've seen Dokku lose steam and then pick up again - but still not be a solution for more than one server.

I love the idea of something that will Just Work, but a lot of stuff doesn't handle the parts I care about the most. Will my one-click PostgreSQL have a couple replicas in my cluster? Will it launch pg-backrest so I can have it backed up to S3? Will it deal with secret storage so I can just add my secrets in the UI and then have my apps grab those secrets as environment vars? Will it fail-over nicely?

Coolify looks really cool and I do want to check it out for real later (probably over the weekend to be realistic). However, it seems like it's single-server (for now) and that seems less compelling for me at the moment. It looks like you're planning on K8s support which would just be wonderful.

Part of me wishes I could get a much simplified UI for K8s. I use Lens at the moment and while it's very good, it doesn't really smooth over the rough bits of K8s, just kinda organizes and presents them. K8s isn't as bad as a lot of people say once you get past the pain of learning. Still, it misses the mark if what you just want a few simple things Heroku-style instead of every bell and whistle. Small users don't care about ingress, they just want their app to be accessible. Small users aren't interested in namespaces or storage classes or roles. Sure, it's important for K8s to support storage classes for large users who might have all sorts of opinions on how they want their data stored. For a small user, you often just want it stored on the local disk.

Likewise, I think there's a reasonably small set of things people often want to run. For example, the databases Coolify supports (MongoDB, MySQL, PostgreSQL, CouchDB, RedisDB) are probably what 95% of people want. There are K8s operators to run them, but like K8s itself it can often be "here's every knob we could think of, we documented 80% of them, now go write some yaml." Heroku lets me deploy PostgreSQL without that. Why can't I get something a bit more slimmed down like Heroku on K8s - so that if I need the additional power in the future, I can get my hands on it.

I'd even love it if this hypothetical K8s Heroku could commit the yaml it is going to apply to a GitHub repository for me so I can easily track infrastructure changes. It would even let people become more comfortable with K8s as they saw the changes they were making in the UI show up as commits in a repository.

I think the point of this rambling is that Coolify is kinda what people want from one perspective - something that seems nice and friendly and handle the cases that the vast majority of people need solved. However, it lacks the critical mass of K8s which can always leave its future in doubt (it looks like there are two of you on this project and open source can take a toll on people) and without high-availability/multi-server it feels a bit lacking for what so many people are looking for in a self-hosted Heroku. Coolify feels like what I'm looking for - I just want it with high-availability and the possibility of continuing on even if the project shuts down (because I can just use an underlying K8s layer, not because it's open source and I could become the new maintainer).

It feels like there's two camps that don't talk to each other: awesome UX people who know what people want and people building solutions that companies want while scaling up (but end up with a bit of a UX mess).

As others have mentioned, the current site (README.md) says very little about what Flynn is.

You can see a version of the repo and README.md, just before it became unmaintained, here:

https://github.com/flynn/flynn/tree/2c20757de8b32a40ba06f7e5...

IMHO it would have been much better if the maintainers had added the "Flynn is Unmaintained" information to the top of the README.md rather than removing all the existing information. I'll submit a GitHub issue suggesting it.

A plea to project maintainers in general: Please include enough information in your top-level README (or README.md, or README.whatever) so that someone who has never heard of your project can find get a good overview. I've also seen READMEs that only discuss the most recent changes. That's fine for readers who have already been following the project, but not for a more general audience.