Kubernetes Tips and News
-
Packaging Job scripts in Kubernetes operators
When using a complex Kubernetes operator, you often have to orchestrate Jobs to perform workload tasks. Examples of Job implementations typically provide trivial scripts written directly in the manifest. In any reasonably-complex application, however, determining how to handle more-than-trivial scripts can be challenging.
In the past, I've tackled this problem by including my scripts in an application image. This approach works well enough, but it does have a drawback. Anytime changes are required, I'm forced to rebuild the application image to include the revisions. This is a lot of time wasted, especially when my application image takes a significant amount of time to build. This also means that I'm maintaining both an application image and an operator image. If my operator repository doesn't include the application image, then I'm making related changes across repositories. Ultimately, I'm multiplying the number of commits I make, and complicating my workflow. Every change means I have to manage and synchronize commits and image references between repositories.
Given these challenges, I wanted to find a way to keep my Job scripts within my operator's code base. This way, I could revise my scripts in tandem with my operator's reconciliation logic. My goal was to devise a workflow that would only require me to rebuild the operator's image when I needed to make revisions to my scripts. Fortunately, I use the Go programming language, which provides the immensely helpful go:embed feature. This allows developers to package text files in with their application's binary. By leveraging this feature, I've found that I can maintain my Job scripts within my operator's image.
-
Kubernetes 1.25: PodHasNetwork condition for pods | Kubernetes
Kubernetes 1.25 introduces Alpha support for a new kubelet-managed pod condition in the status field of a pod: PodHasNetwork. The kubelet, for a worker node, will use the PodHasNetwork condition to accurately surface the initialization state of a pod from the perspective of pod sandbox creation and network configuration by a container runtime (typically in coordination with CNI plugins). The kubelet starts to pull container images and start individual containers (including init containers) after the status of the PodHasNetwork condition is set to True. Metrics collection services that report latency of pod initialization from a cluster infrastructural perspective (i.e. agnostic of per container characteristics like image size or payload) can utilize the PodHasNetwork condition to accurately generate Service Level Indicators (SLIs). Certain operators or controllers that manage underlying pods may utilize the PodHasNetwork condition to optimize the set of actions performed when pods repeatedly fail to come up.