If you’re a Docker Cloud user, you recently woke up to some disconcerting news regarding the longevity of the Docker Cloud platform. Docker announced that it would be discontinuing support for cluster management in their Docker Cloud product on May 21st, in turn leaving many users scrambling to find an alternative. If you’re not familiar with the product, Docker Cloud was born out of the 2015 acquisition of Tutum, and provides users with a simple SaaS interface to build, ship, and run Docker containers. Tutum (and subsequently Docker Cloud) was one of the first mature SaaS platforms for container orchestration in the market. At the time, developers were struggling to take Docker images running locally, and move them from a development environment to a production infrastructure which required orchestration across an indeterminate number of machines. Given that Docker owned the underlying image format and runtime, providing its users with orchestration was a logical next step. Naturally, after the acquisition of Tutum, Docker integrated its Swarm orchestration into Docker Cloud, building a solid foundation for running Docker containers in production.
The War is Over
But the fact is, it’s no longer 2015; Swarm is no longer the major player. While researching replacements for Docker Cloud, organizations should carefully consider the current orchestration landscape. In just two short years, Kubernetes has skyrocketed to the top of the orchestration solution list; in fact, many would argue that it’s hard to find a close second.
After a short and nearly uncontested fight, Kubernetes won the proverbial orchestration war. Users can (and will) continue to quarrel about the best overall platform on the market, but the fact is: all leading platforms are built on Kubernetes. Although late to the game, Docker is now accepting defeat, offering support for Kubernetes in both Docker Community Edition and Docker Enterprise Edition. So what does this mean for current Docker Cloud users?
Put simply: don’t walk, run to Kubernetes.
The Learning Curve
Not dissimilar to Docker, getting started with Kubernetes can be a bit daunting. While the majority of core Kubernetes concepts are ultimately trivial to grok, there is a wealth of terminology users must first master. Drain, cordon, taints, tolerances, pod disruption budgets, daemonsets, replicasets; the list goes on. Learning the vocabulary is necessary just to set the table. Keep in mind, this is on top of core containerization and infrastructure knowledge users must have. Once foundational knowledge is acquired, users must then take on the challenge of getting a cluster stood up, applications migrated, role-based access control in place, monitoring, alerting and log collection integrated. This is in addition to all of the other complexities that come with running production infrastructure. Included in the official migration guide produced by Docker is a Kubernetes primer which provides a high-level overview of Kubernetes concepts. Other great resources to help newcomers come up to speed are Katacoda and Kubernetes the Hard Way. When coming from a plug and play SaaS solution, it is easy to get overwhelmed by the number of concepts included in Kubernetes. But remember, Kubernetes provides low-level building blocks; solution providers have already gone through the laborious task of building platforms on top of Kubernetes, making it easier to manage infrastructure. As an example, Containership provides log streaming through integration with the Kubernetes API and metric collection through the popular Cloud Native Computing Foundation project, Prometheus.
Also included in Docker’s official migration guide are walkthroughs for migration to Azure Kubernetes Service and Google Kubernetes Engine. Each guide contains a section on migrating your existing Cloud Stack (a Docker Cloud concept) to the standardized Kubernetes manifest, which can be used on any Kubernetes platform abiding by the API standards of upstream Kubernetes. As more cloud providers support the CNCF, native support for Kubernetes will follow. Since providers will provide CNCF certified platforms, interoperability and portability will be inherent. The breadth of Kubernetes adoption unlocks the foundation for multi and hybrid cloud infrastructure. Being stuck in an ecosystem that only supports Swarm not only restricts flexibility, but stunts innovation. Imagine using the same deployment manifests to stand up a development environment on your local machine, a staging environment in Digital Ocean, and production environments spread across Google Cloud Platform and Amazon Web Services, in multiple regions. And by leveraging a Kubernetes management platform such as Containership Cloud, managing workloads and resources across clusters becomes trivial.
At the time of writing this, you have less than two months left to find a new platform. Why not try Containership? Containership provides an enterprise ready and developer friendly UI for managing Kubernetes clusters and workloads. Our team is offering free migrations from Docker Cloud to Containership Cloud with new annual managed support contracts.
Contact us if you’re interested in learning more about how Containership can help migrate your infrastructure from Docker Cloud to a managed Kubernetes solution!