How do you observe a wildly standard undertaking like Kubernetes? How do you determine the place it’s going? If you’re contributing to the undertaking or taking part in Special Interest Groups (SIGs), you may acquire perception by osmosis, however for these of you with day jobs that don’t embrace contributing to Kubernetes, you may like somewhat assist studying the tea leaves. With a fast-moving undertaking like Kubernetes, the top of the 12 months is a superb time to check out the previous 12 months to realize perception into the following one.
This 12 months, Kubernetes made quite a lot of progress. Aside from inspecting code, documentation, and assembly notes, one other good supply is weblog entries. To acquire some insights, I took a take a look at the highest ten Kubernetes articles on Opensource.com. These articles give us perception into what matters individuals are curious about studying, however simply as importantly, what articles individuals are curious about writing. Let’s dig in!
(Get the total listing of high 10 Kubernetes articles from 2019 on the finish.)
First, I’d level out that 5 of those articles deal with the enlargement of workloads and the place they will run. This enlargement of workloads consists of information science, PostgreSQL, InfluxDB, and Grafana (as a workload, not simply to observe the cluster itself) and Edge. Historically, Kubernetes and containers typically have largely run on high of digital machines, particularly when run on infrastructure supplied by cloud suppliers. With this curiosity in Kubernetes on the edge, it’s one other signal that finish customers are genuinely curious about Kubernetes on naked steel (see additionally Kubernetes on metal with OpenShift).
Next, there appears to be quite a lot of starvation for operational information and finest practices with Kubernetes. From Kubernetes Operators, to Kubernetes Controllers, from Secrets to ConfigMaps, builders and operators alike are on the lookout for finest practices and methods to simplify workload deployment and administration. Often we get caught up within the precise configuration instance, or how folks do it, and don’t take a step again to appreciate that each one of those fall into the bucket of how one can operationalize the deployment of purposes (not how one can set up or run Kubernetes itself).
Finally, folks appear to be actually curious about getting began. In reality, there may be a lot data on how one can construct Kubernetes that it intimidates folks and will get them down the improper path. A few the highest articles deal with why you need to study to run purposes on Kubernetes as an alternative of concentrating on putting in it. Like finest practices, folks usually don’t take a step again to investigate the place they need to make investments their time when getting began. I’ve all the time advocated for, the place attainable, spending restricted money and time on utilizing expertise as an alternative of constructing it.
5 predictions for Kubernetes in 2020
So, trying again at these themes from 2019, what does this inform us about the place 2020 goes? Well, combining perception from these articles with my very own broad purview, I need to share my ideas for 2020 and past:
Expansion of workloads. I’d preserve my eye on high-performance computing, AI/ML, and stateful workloads utilizing Operators.
- More concrete finest practices, particularly round mature requirements like PCI, HIPAA, NIST, and many others.
- Increased safety round rootless and better safety runtimes classes (like gVisor, Kata Containers, and many others.)
- Better standardization on Kubernetes manifests because the core artifact for deployment in improvement and sharing purposes between builders. Things like podman generate kube, podman play kube, and multi functional Kubernetes environments like CodeReady Containers (CRC)
- An ever-wider ecosystem of community, storage and specialised (GPUs, and many others.) distributors creating better of breed options for Kubernetes (in free software program, we consider that open ecosystems are higher than vertically built-in options)
I am trying ahead to a different nice 12 months in Kubernetes!
Top 10 Kubernetes articles for 2019