Virtualization Technology News and Information
StormForge 2024 Predictions: Reducing the cognitive load of Kubernetes


Industry executives and experts share their predictions for 2024.  Read them in this 16th annual series exclusive.

Reducing the cognitive load of Kubernetes

By Yasmin Rajabi, VP of Product, StormForge

At this point, Kubernetes has become the de facto platform for enterprises deploying modern software; KubeCon + CloudNativeCon North America 2023 was a great testament to that. From small startups to large banks, most folks have the majority of their infrastructure running on Kubernetes and have been spending the last year focused on the toil that is day 2 scaling challenges. As we look to next year, their focus will narrow to how they can stop spending so much time dealing with Kubernetes management and more about pushing the business forward. Specifically, infrastructure teams will spend more effort on how they can reduce the cognitive load on their developers and their platform teams, focusing on what investments they need to make to free up engineers to invest time on business initiatives and less on managing the infrastructure. With the large size Kubernetes deployments across many different industries, 2024 will be all about reducing the infrastructure burden on engineers and freeing up engineering time to focus on what matters.

Kubernetes Resource Management

In 2024, Kubernetes resource management will no longer be an afterthought that causes performance and cost problems. Teams will realize the monotonous activity of trying to set requests, or better yet, not setting them at all and running unbounded, is not only risky and costly for the company, but also a waste of an engineer's time. Given the deployment size of Kubernetes at organizations, in the next year setting requests correctly/optimally will be an unavoidable focus, with acknowledgement that it's a challenge worth tackling. Providing solutions for proper resource management will be a topic that hits the priority list of platform engineers.

High Frequency Automation

It seems like all predictions include "more automation", especially when we are talking about infrastructure management. However, surprisingly, most folks are still hesitant to automate away the toil when it comes to resource management in Kubernetes. Practices like GitOps help, but when you are trying to continuously rightsize at the scale of the enterprise, think 100,000s of workloads, robust automation practices that meet the needs of platform engineers need to be implemented. Just deploying changes via automation doesn't cut it. Automating something once a week leaves a lot of opportunity for change and drift. Organizations will look to high frequency automation that can continuously detect a need for change and address it without the need for human intervention. 

Machine Learning

2023 was the year of ChatGPT, and 2024 will be the year infrastructure teams adopt more machine learning and AI into their workflows. What was once an interesting idea approached with trepidation will become the standard. Every changing dataset, with hourly, weekly, and monthly trends in application behavior drive a clear need for machine learning. Scaling behavior, how workloads interact with one another, and how often different node types are available are all perfect problems for machine learning. The ability to learn infrastructure patterns and predict trends that can easily be accessible with a simple question will be something platform teams will look to adopt to keep up with the pace of innovation and keep their internal customers happy.

As we wrap up 2023 and look to the next year, I have high confidence in organizations' ability to mature their Kubernetes practices and reduce the burden on the users. It won't be easy at first, but I feel strongly that we have hit a scale point with Kubernetes that requires engineers to take a step back and ask if the toil and cognitive load of their day to day battle with Kubernetes is really the best way. They will stop trying to fight with resource management and look to automation and machine learning to reduce their cognitive load.



Yasmin Rajabi, VP of Product, StormForge

Yasmin Rajabi 

Yasmin is a product leader focused on building products that empower engineers to move the business forward. As VP of Product at StormForge, she is responsible for Product Management, UX, and Docs. Previously Yasmin was Senior Director of Product at Puppet, leading teams across multiple products in the portfolio. Prior to that, she was a Site Reliability Engineering Manager at Staples.

Published Monday, December 04, 2023 7:33 AM by David Marshall
There are no comments for this post.
To post a comment, you must be a registered user. Registration is free and easy! Sign up now!
<December 2023>