Akuity this week at the KubeCon + CloudNativeCon Europe conference revealed it has added an ability to customize the steps used to promote applications into a production environment using a Kargo orchestration engine it developed to manage software using a GitOps workflow.
Company CEO Hong Wang said the Custom Steps capability added to Kargo will enable software engineering teams to define any promotion logic as a native step in a pipeline, including, for example, running a policy check or security scan.
The overall goal is to eliminate the need to create custom scripts or rely on manual processes to extend a GitOps workflow that is being used to move container applications into a production environment, he added.
Custom Steps adds a container-based interface for defining promotion logic. Each step runs in a Kubernetes Pod at execution time, with its inputs, outputs, and result recorded in the Kargo promotion record alongside every other step in the sequence. Custom Steps are packaged as OCI images and registered to enable any team to reference a custom step by name in their promotion templates.
Originally built for Kubernetes as a complement to the Argo continuous delivery (CD) platform, Kargo is now being used widely to promote software across a range of platforms, said Wang. It includes a library of built-in promotion steps covering the most common GitOps promotion operations, including Argo CD syncing, Helm and Kustomize rendering, and Terraform infrastructure management.
Since the initial launch of Kargo, there have now been more than 3.5 million downloads, with Deutsche Telekom, JumpCloud and Cisco ThousandEye among the early adopters.
It’s not clear how many DevOps teams are moving towards a GitOps workflow that loosely couples continuous integration (CI) and CD workflows. Historically, DevOps teams have adopted CI/CD platforms to unify the management of those workflows, but as Kubernetes clusters become more widely deployed many organizations have adopted ArgoCD to separately manage CD workflows via a graphical user interface (GUI). That approach provides the added benefit of making Kubernetes environments more accessible via a higher level of abstraction.
Kargo, meanwhile, provides an orchestration layer that can be more widely used to manage the pipelines that DevOps teams create to deploy software in a way that is not tied to a specific platform.
Each DevOps team will need to consider how they might want to approach CD going forward. While many CI processes are automated, the percentage of DevOps teams that have automated CD is much smaller, mainly because each platform for deploying application software has a unique set of attributes. However, with the advent of orchestration layers such as Kargo there may be an opportunity to revisit how applications are deployed in production environments.
Regardless of approach, the one thing that is certain is that as the volume of code being created using artificial intelligence (AI) tools continues to increase the issue one way or another will soon be forced. The issue, as always, is overcoming the current level of inertia that tends to surround any legacy approach to building and deploying applications.