Toby Bellwood ·
May 14, 2020 · 3 min read
Update July 2020
In a continuing effort to increase our transparency and communication, we are starting to document more about our infrastructure setup and configuration. We realise that there are a lot of Managed Kubernetes providers out there (the list at the CNCF Certified Hosted Kubernetes Landscape is ever-growing), so we will endeavour to outline how we go about evaluating a platform for use with Lagoon. The majority of our decision-making process here is customer/partner driven, but there are a number of criteria that we use to evaluate the services offered by the different providers, and we’d like you to understand why they’re important to us, and what they mean.
We have started a discussion page over at our GitHub repository to capture our thinking, which can also be used to capture any comments/discussion. Head on over to https://github.com/amazeeio/lagoon/discussions/2035 to watch it unfold!
In our recent blog post Michael outlined why amazee.io is working so hard to make Lagoon work in managed Kubernetes. The benefits of better autoscaling and reduced maintenance obviously are the headline grabbers, but today we want to discuss one of the most important benefits to amazee.io and our customers: portability.
The team has been hard at work with a number of managed Kubernetes providers, already deploying sites into AWS Elastic Kubernetes Service (EKS) in three regions, and Azure Kubernetes Service (AKS) in three more, as well as building a relationship with Catalyst Cloud in New Zealand. Ensuring that we are familiar with the products, processes, and nuances of the various providers is essential to our success.
The widening of the infrastructure choice has helped us reduce a major barrier to entry for our customers - they can utilise existing hosting relationships (billing, accounts, permissions etc) to bring their sites to Lagoon. The other opportunity this realises is the ability to host your sites “closer to home.” Where previously Lagoon was mostly limited to the large amazee.io clusters in the three main regions, now with a wide range of EKS, AKS and GKE (Google Cloud) locations available, you’ll be spoilt for choice!
We’ve developed a process that can handle not only the Openshift to Kubernetes migrations, but also migrations between different Kubernetes clusters (even across different providers and regions!) Above all, amazee.io can manage the migration for you, with no downtime, and no-need to re-tool your development environments!
We will be reaching out to all our existing customers to schedule a switch over the coming months, but if you have specific requirements or timeframes, please let us know. When it’s time to go, we will migrate all your environments from one cluster to the other, give you a dedicated url to test your sites on the new cluster, and when all looks good, we do one last sync and switch the DNS. Migrations in a breeze!
With the ability to migrate projects between a customer’s Kubernetes clusters as they deploy more sites, we can now work closely with them to find the optimum cluster architecture -- looking closely at node and pod counts, anticipated workloads, and traffic patterns. Individual clusters could be provisioned for segregated workloads, the provisioning of development environments, or configured for occasional on-demand use, so the ability to “move” projects and sites between them will be hugely useful in optimising performance and price.
If you’re interested in hearing more about our Kubernetes offering, check out our infographic.