Weaveworks Updates GitOps Platform for Kubernetes Clusters
Weaveworks Inc. today announced it has updated the Weave Kubernetes Platform (WKP) to enable IT teams to run the GitOps platform for deploying applications on any Kubernetes cluster without any reprovisioning required. Previously, WKP was only available on a distribution of Kubernetes from Weaveworks.
Version 2.4 of WKP enables this capability by adding support for a cluster lifecycle management API (CAPI) that provides a standard declarative cluster definition format and a set of installers and reconciliation agents for Kubernetes clusters.
This latest release also includes support for Cluster API for Existing Infrastructure (CAPEI) that enables IT teams to leverage tools such as Terraform to install Kubernetes cluster management tools on pre-provisioned machines without requiring a central management cluster.
Weaveworks COO Steve George says that approach will make it easier to deploy management tools on Kubernetes clusters deployed on edge computing platforms.
Finally, WKP 2.4 also adds multi-tenant support to Team Workspaces, a workflow application for tracking changes to Git-based deployments that can now be used by multiple DevOps teams. Each workspace can also span multiple Kubernetes clusters to simplify the rollout of applications across a fleet of Kubernetes clusters.
George says WKP enables IT organizations that have embraced Kubernetes to transition to a GitOps-based approach to application deployment that enables code to be pulled automatically from a Git repository when it’s ready. Most existing DevOps platforms in contrast enable IT teams to manually push code from a repository onto a specific platform. That approach is required because historically each platform the code might have been deployed on was unique. Kubernetes provides a layer of abstraction that is consistent across multiple platforms, which then makes it possible to automate the process of pulling code out of a repository whenever it’s available.
Weaveworks takes the concept to the next level by also automating the deployment of Kubernetes clusters along with complementary modules such as the open source Prometheus monitoring tools, he adds.
Transitioning to a GitOps-based approach to continuous delivery (CD) obviously becomes much easier when Kubernetes clusters are present. The challenge is, most organizations today are still deploying code on legacy platforms. As such, many may find themselves supporting multiple approaches to CD for many years to come. Most new cloud-native applications, however, are being constructed using microservices based on containers that are generally deployed on Kubernetes clusters.
It may take a while for organizations to transition to GitOps-based workflows for deploying applications. However, given the fact that most organizations that have embraced best DevOps practices have tended to limit their focus to continuous integration (CI) during the application development process, it’s clear there is much opportunity to revisit CD. In fact, some organizations may go so far as to further separate CI and CD into two distinct functions managed by separate platforms.
Whatever the path chosen forward, the one thing that is clear is best DevOps practices are about to evolve once again.
