We have created ContinuousPipe
3 years ago, I was starting to push the first commits to a side project aiming at simplifying the deployment of my Docker containers (on a Kubernetes 0.x cluster at the time).
A year and a half later, my team and many others at Inviqa were using this project called ContinuousPipe as the main tool to deploy our “Dockerized” applications to Kubernetes clusters running on AWS and Google Cloud. It clearly demonstrated its value:
- Low-cost isolated environments. Every in-development feature has its own environment (each “pull-requests”) at a very low cost, as those are now “shared” on a single Kubernetes cluster.
- New projects and services on-boarded very fast. New applications’ continuous deployment configuration are drastically simplified from an operational perspective, as all it requires is a Docker configuration and a Git repository.
- Very flexible. Via its filters and ability to use labels on pull-requests, ContinuousPipe allows to customise for very specific and hectic use cases.
Following an internal success within Inviqa, we decided to take ContinuousPipe to the next level and opened it to a wider audience as a SaaS product.
We have learnt so much!
This was fantastic adventure. Something I couldn’t have done my myself and I’ve been lucky to work with fantastic and incredible people always supporting me with great ideas and recommendation to improve ContinuousPipe. We learnt so much more about the technologies (Docker and Kubernetes especially — and their nice “itches”), the processes (branch vs trunk based, the “operations” side of the things, etc…) but also the “market”.
The area of “containers” or more broadly the answer to “I need to have my application online” is changing constantly. I’m not even sure when it really started, but to me AWS have thrown the first massive stone, after what Docker was created, and then Kubernetes did. All of which proposing variety of great tools to create containers, however I found there was a sense of complexity with that respect which was part of the reason ContinuousPipe was created, to make what was complex simple giving me more time to focus on more valuable things. On top of such tools, platforms like Heroku are improving the life of million of developers by massively removing most of the complexity, and now, there is an impressive force dragging infrastructure to “serverless”, which ultimately just means there is no concept of “VM” or “container” anymore.
Such volatility is great for the innovation but forces actors of this field to invest an insane amount of money to keep up with the pace of change while being profitable as an organization. Looking at this from a different angle those new ways of doing things, actually lead to new needs from the end users, the developers who, let’s be honest, are not the easiest customer: they want it now, free and including features they’ll never use.
Ultimately, all we want is to increase the ability for developers to focus on what brings value to the business rather than spending hours on packaging and deploying their applications.This is all what ContinuousPipe is about, making the complex simple and give valued time back to the developers.
We are open-sourcing it
We had two choices: scale with a massive external investment or closing the doors of the SaaS version of the product. We took the second one. In order to continue with our mission of helping as much as possible development teams, we decided to open-source the entire codebase.
Now, you can browse the source-code and get started with it on your own Kubernetes cluster! We hope to be able to leverage these years of hard work for the best and see you in our Slack.
As always, with 💖 from Sam.