Rancher 2.5 Embraces GitOps at Scale with Rancher Continuous Delivery

Rancher 2.5 Embraces GitOps at Scale with Rancher Continuous Delivery

Keith Basil
Keith Basil
Gray Calendar Icon Published: October 6, 2020
Gray Calendar Icon Updated: October 20, 2020
Discover what’s new in Rancher 2.5

The ability of Kubernetes to easily deploy and manage containerized software has given organizations tremendous capabilities in their cloud services, with clusters multiplying into the hundreds or thousands and extending out to the edge for any number of purposes. But its growing popularity has also led to challenges in managing complexity in an environment that is conducive to cluster sprawl.

Rancher 2.5 simplifies deployment and management with Rancher Continuous Delivery, a new capability that adopts a GitOps approach toward streamlining the process across any infrastructure—private, public, hybrid or multi-cloud environments—by using a Git repository to store and manage application and configuration information automatically, ensuring consistency in deployments while taking a lot of the load off of the customer. The Fleet project, introduced earlier this year, is the engine behind Rancher Continuous Delivery. Fleet is a Kubernetes cluster controller designed to address the challenges of running thousands to millions of clusters worldwide.

K3s Takes Kubernetes to the Edge

As Kubernetes adoption grows, organizations have steadily increased the number of clusters they deploy, and with Rancher's lightweight K3s have taken thousands of standalone clusters to the edge. K3s greatly streamlines deployment at the edge by eliminating the complexity and cognitive overload of trying to install Kubernetes. Instead of a detailed manual process, you can, for example, use a single Linux command to get a one-node Kubernetes cluster running, ready for development and testing as a full blown, CNCF-certified distribution of Kubernetes—all in about 45 seconds. K3s is architecture-independent and has an extremely small footprint. It will run on ARM or Intel processors, without taking up resources. A \$40 Raspberry Pi will do the job.

That simplicity allows organizations to create value higher up into the stack, fueling the momentum for edge deployments and enabling a true digital transformation – with customers deploying to hundreds or even thousands of locations. The average number of clusters in one sample set, for instance, has been about 1,700 remote locations. That gives them an expansive reach into the cloud, but managing 1,700 of anything is difficult. That's where the Rancher 2.5 Continuous Delivery capability comes in.

Rancher Continuous Delivery Uses Git Developer Tools

The underlying GitOps approach builds on mature and well-known workflow processes understood by both developer and operations personnel. Rancher Continuous Delivery leverages existing Git developer toolsets for managing cloud-native applications, ensuring that the applications run as intended. A Git repository is the sole source of authority for the configuration. All definitions are held at the source, whether they are Kubernetes cluster configuration definitions, application deployment definitions or, in the near future, security policy definitions. When developers make changes to an application, Git can push it out to Kubernetes for deployment. If changes deviate from the desired state, Git can put them back in line.

Git typically works on a push or pull method, with changes being sent directly (push) and delivered after clusters send a periodic request for updates. Rancher 2.5 goes further with the pull method. Rather than going directly to Git, changes go to its centralized Fleet controller, which manages access to the Git repository as well as the configurations and the assignments between the repository and the applications or clusters, so that the right code gets to the right application. Within Rancher 2.5, the Rancher Continuous Delivery feature provides an effective model for managing things at scale. Application packages can be associated with an individual cluster or a group of clusters, up to one million.

CONTINUOUS DELIVERY CHART

In fact, scaling the number of clusters you manage requires shifting your effort and attention away from managing each individual cluster. Just like Kubernetes has shifted the focus away from individual computing nodes, we want to do the same with clusters themselves. The above diagram shows how Rancher 2.5 lets you define bundles that target clusters based on selectors. We refer to the units as bundles and not applications because not only can Rancher 2.5 manage application deployments, it can handle any Kubernetes resource. Rancher Continuous Delivery provides a built-in mechanism to customize bundles per target cluster using industry-standard tools such as Helm and Kustomize. Once bundles are deployed across clusters, Rancher 2.5 actively monitors that resources are ready, consistent and secure. The primary use case is to ensure that deployments are consistent across clusters. You can deploy applications or easily enforce standards such as "every cluster must have X security tool installed."

Organizations want to run Kubernetes and containers at the edge, but they don't want to take on the burden of managing the things below Kubernetes. Rancher Continuous Delivery in Rancher 2.5 gives customers a controller that allows them to manage the unique requirements for Kubernetes on the edge, and do it at scale, regardless of platform. It works with Azure, Amazon, Google or a bare-metal instance on prem.

Rancher Continuous Delivery Extends CI/CD Workflows

Rancher Continuous Delivery works hand-in-hand and extends continuous integration/continuous delivery (CI/CD) workflows, where developers constantly build and deploy applications, allowing control of new applications and updates as they appear. This new capability in Rancher 2.5 delivers full lifecycle management for thousands of Kubernetes clusters at the speed of DevOps software development.

The explosion in cloud services has created a growing number and increasing diversity of clusters out there in the wild. Rancher 2.5, which can be deployed on any CNCF-certified Kubernetes distribution at any scale, provides supreme command and control over what's running on those clusters, as well as the actual cluster configuration. Rancher Continuous Delivery's GitOps approach, being declarative in nature, matches up well with Kubernetes in making sure resources are where they're supposed to be. It represents the next step in the evolution of distributed container management. By streamlining application delivery across any infrastructure in any location, enterprises can use Rancher 2.5 to accelerate their journey toward true digital transformation.

Discover what’s new in Rancher 2.5
Keith Basil
Keith Basil
VP, Edge Solutions
Keith leads the strategy and management of Rancher’s edge products such as K3s and Fleet. He works with Rancher’s global customer base on Kubernetes-centered edge solutions that encompass cluster management, heterogenous architectures and zero-trust security approaches at scale.
Get started with Rancher