Kubernetes has a bad habit of throttling CPU resources—leading to severely degraded performance or spiralling costs for extra, unnecessary infrastructure.
We put a K8s cluster in the lab to monitor how and when CPU throttling happens. Then we tried to kill the cluster by overloading it with CPU-hungry containers.
Watch this video to see how K8s clusters protect themselves from what they see as heavy CPU usage, and how you can monitor and troubleshoot the problem.
We demonstrate how you can:
– Use Netdata to reduce API response times by a factor of 7
– Expect to reduce infrastructure resource requirements by 60-75%
00:00 Introduction
11:03 To Throttle or Not to Throttle
24:06 Kill the Cluster
43:07 Takeaways
Want to know more? Hit subscribe now to get the next installment of our series on Kubernetes monitoring and troubleshooting, then visit:
Our website: https://www.netdata.cloud/integrations/kubernetes-monitoring/
Our documentation: https://learn.netdata.cloud/guides/monitor/kubernetes-k8s-netdata
Thoughts, questions or suggestions?
Share them on our Discord server: https://discord.gg/4ShKb65VvF
Or on our community forum: https://community.netdata.cloud/