Komodor: LIVE Workshop: Hidden Signals in K8s Clusters: A Data-Driven Approach to Reliability

What can we learn from observing Kubernetes clusters in the wild, and analyzing their behavioral patterns? Which hidden signals are we missing?

At Komodor, we saw an opportunity to leverage the troves of data hiding within Kubernetes clusters to help users uncover and resolve complex problems stemming from hidden sources that are not trivial and immediately obvious.

Through our "Reliability Insights" project, we aimed to analyze raw data, generated from observing hundreds of Kubernetes clusters in the wild, and transform it into actionable intelligence.

Extensive research and experimentation led us to develop methods to clean and process this data, uncovering remarkable findings. We identified multiple categories of reliability-related insights, each offering a unique perspective on cluster health and performance, such as workloads with CPU throttled too much, impact of Node termination due to SPOT instance usage, and more!

Join our very own Andrei to learn about our journey of discovery, and:

✅ Explore how we combined various data points to reveal hidden issues

✅ Discuss the challenges we faced in making sense of Kubernetes' vast data landscape

✅ Discover how these insights can level up your cluster reliability management

Register for free to save your spot!