We decided to build gaudi 10 months ago because we believe that containers are the missing enabler for a true microservice architecture. But Docker alone isn’t enough to setup a multi-container environment: linking is complicated, sharing resources isn’t easy, and storing a lot of Dockerfiles doesn’t make it. So we developed a container orchestrator for Docker, in Go. When I say we, it’s mostly a one man work: Emmanuel Quentin really deserves all the praises on this project.
But on July this year, Docker acquired Orchard. The company is mostly known for their orchestration tool called Fig. That’s right, Docker invested in an orchestration solution for Linux containers. It’s a great move by Docker, because it makes their offer more complete, and allows them to address an existing requirement without redeveloping a solution from scratch.
Fig and gaudi address the same needs. In fact, 10 months ago, Fig and gaudi were very similar. Since then, Orchard kept on investing in Fig, while marmelab didn’t invest much in gaudi. Sure, we have a website, a demo and a few blog posts, but that doesn’t make a product.
Today, Fig is the de facto standard for Docker orchestration. It is much more complete than gaudi, it has strong traction by Docker. It’s a great tool - we even use it ourselves!
It doesn’t make sense to keep on developing our own solution anymore. Today, marmelab is officially retiring gaudi. Being agile also means acknowledging failure, and having the courage to stop a project. 525 stars on GitHub. About 9,000 unique visitors. 6 months of development. That’s the price of the lesson we learned: if there is an opportunity, invest in it, take a product strategy, make it visible, use it extensively, solve people’s problems. We can’t be happy just because we learned Go and Docker. Making successful products is what we like.
I want to thank all the early adopters who tried gaudi and sent us their feedback. I want to thank all the current gaudi users, and tell them that they will find a great alternative in Fig.
But don’t worry for us: we have lots of ideas for future products, and a few projects already in good shape (ng-admin, phpcr-browser, gremlins.js). Keep on using our products, and if you like them, share the love!