Mon Petit Placement is a French startup specializing in investing. As they scaled, they needed an infrastructure to support their growth, so they teamed up with Skyloud. Skyloud supported them in migrating their infrastructure to the cloud, and more particularly on Kubernetes. I interviewed Skyloud’s CTO, Etienne Fachaux, to understand how they used Kubernetes and Terraform to migrate VM-based infrastructure to the cloud.
The migration came from the need to have a scalable platform to do FinOps. Mon Petit Placement needed the cost of their infrastructure to be optimized and based on their actual use: as nothing happens at night, and they had a big spike during the day, they didn’t want to pay for fixed resources.
The original architecture was based on dedicated servers and instances, which required too much time to install and configure, but mostly blocked the flexibility and scalability they needed. It was neither scalable nor redundant, as they had a single point of failure. So if there’s a down, they had no backup solution except reinstalling the service.
So they called Skyloud to build new infrastructure on the Cloud to unlock the cloud’s benefits.
The most challenging part of migrating to the Cloud is restructuring the application to make it cloud-native and easily usable for the developers. So they did it in two steps: one to educate on the new practices and new deployment methods engineers on Mon Petit Placement’s team, and the second to migrate, install and test everything.
This allowed Mon Petit Placement’s engineering team to have time to get onboarded on those drastic changes.
Mon Petit Placement and Skyloud chose Kubernetes as the main component of their infrastructure, with complementary solutions such as a personalized CI/CD. The migration was smooth thanks to Terraform and Kapsule, which allowed Skyloud to test and deploy quickly.
Terraform was a great support in executing this project: it was the obvious choice to have every resource documented on its implementation in a coded form. Terraform helped give full transparency on how their infrastructure got deployed on the cloud. That enabled Mon Petit Placement to evolve without being stuck and needing to call on Skyloud at every step. They only had benefits thanks to this migration.
Kubernetes has a strong community that brings a lot to the cloud industry and has the benefits of being cloud-agnostic and open-source. All those reasons make Kubernetes particularly loved at Skyloud - which is also a significant open-source contributor.
Kubernetes also enables its clients to switch to a multi-cloud strategy whenever they want to, thanks to the infrastructure delivered by Skyloud on Kubernetes. Skyloud provides an entirely done infrastructure, with a door open to the multi-cloud, to let the company be free to choose and be flexible on their growth. They don’t want their client to depend on them, or a specific cloud provider - which means providing the best tools in an agile environment.
They chose Kubernetes to implement this new infrastructure as it meets the fundamental principles of good infrastructure and applicative architecture. The standardization of Kubernetes brings flexibility, agility and enables the developers to test components that were inaccessible before Kubernetes.
Kubernetes allows interfacing two clouds thanks to the compatible connector. If clouds don’t work together on shared projects to unify the cloud industry, then their client pays the price; it’s complicated for us to have interoperability between clouds to allow secure communication between clouds.
There is a strong need for companies to control the flow between clouds.
For Skyloud, Kubernetes is a response to that, or Rancher too. Rancher is an open-source tool we particularly enjoy thanks to its polyvalence and its interoperability between clouds, which gives more flexibility in our deployment and allows us to benefit from each cloud's strengths.
Significant changes always ask for a pedagogical approach. Without it, a new solution risk not sticking and being used. So Skyloud put in place tools and visuals to support the engineering team of their clients in understanding how their new infrastructure works, all the new deployment methods, and the new CI/CD.
The goal is not to have clients who depend on Skyloud’s service but to empower them. They can intervene in any bricks of the infrastructure, whether it’s Gitlab or Terraform, and add what they want to weed this new infrastructure with new tools and application concepts.
Skyloud handles all the redundant parts and automatizable to provide a clear and finished CI. Then the client adds on whatever he may need - with Skyloud being on their corner if they need support.
So if you are a startup that wants to migrate to the cloud to optimize its infrastructure, with an engineering team to focus on the product to tackle this mission, how about relying on a partner?
Some features may be hard to find and used only by most advanced users. We thought you might be interested in going further and optimizing your use of Kubernetes Kapsule
We will be using a Kubernetes Kosmos cluster, and deep-diving into specific multi-cloud Kubernetes concepts
Kubernetes provides a series of features to ensure your clusters have the right size to handle any load. Let's look into the different auto-scaling tools and learn the difference between them.
