Introduction: Over the past few months, I've shared my insights on the persistent issue of an idle k3s cluster exhibiting escalating CPU usage. In my previous blog posts (https://canthonyscott.com/k3s-cpu-usage-bug-remains/ and https://canthonyscott.com/high-idle-cpu-usage-of-k3s-continuously-increases/), I detailed my observations across k3s versions 1.21, 1.22, and even a follow-up test on version 1.25. Today, I bring good news - it seems like this long-standing bug has finally met its match.
The Resolution: After running a k3s instance of version 1.28 for several days, I am delighted to report that the consistent CPU usage bug appears to be definitively resolved. This marks a significant milestone for the stability of k3s, and I'm eager to share this positive development with the community.
This resolution brings a sigh of relief as I've encountered this bug across multiple versions, and its persistence had raised concerns about deploying k3s in more production-oriented settings. With version 1.28 showing promise, I am now confident and ready to roll out k3s to some of my more critical production environments.
The journey to stability in container orchestration can be fraught with challenges, but it's reassuring to witness the k3s community actively addressing and resolving issues. The timely resolution of the high idle CPU usage bug in version 1.28 signifies progress and instills confidence in the reliability of k3s for diverse use cases.