Discover how eBPF is revolutionizing Kubernetes observability and security and learn how to overcome its performance challenges in this deep dive episode of Observe & Resolve. If you’re exploring eBPF (Extended Berkeley Packet Filter) and its potential in Kubernetes environments, this video is a must-watch. Our experts break down complex concepts to guide you through monitoring and optimizing eBPF resource usage effectively. Here’s what you’ll get:
- A clear explanation of eBPF and its role in kernel-level observability.
- Insights into why Kubernetes’ kubelet might fail to spot critical resource issues.
- Practical demonstrations of using Inspektor Gadget, OpenPipeline, and the Dynatrace Operator for efficient monitoring.
- Two powerful solutions for better observability:
1. Bpfstats for streamlined metric collection.
2. Top eBPF (bpftop) for granular, program-level insights using logs.
We’ll also show you how to set up observability pipelines, transform metrics, and create dashboards to visualize CPU and memory usage of eBPF probes.
By the end of the video, you’ll have actionable tools and strategies to fine-tune your cluster’s performance while maintaining top-tier security.
---------------------------
📖 Chapters 📖
00:00 Introduction
00:30 What is eBPF
00:56 The problem with eBPF and Kubernetes
01:30 Solution: eBPF monitoring
02:01 One small note
02:38 Observability Setup
03:00 Solution 1: bpfstats
03:36 bpfstats: OpenTelemetry Collector
03:58 bpfstats: OpenPipeline
04:18 bpfstats: Dashboard
04:39 Solution 2: bpftop
05:00 bpftop: OpenTelemetry Collector
05:10 bpftop: OpenPipeline
06:06 bpftop: Dashboard
06:22 Conclusion
-----------------------------
🔗 Additional Links
Check out Inspektor Gadget
Visit the repository related to this video on GitHub
Try out Dynatrace with a free trial
-----------------------------
Subscribe to our YT channel
Stay up-to-date with Dynatrace! Follow us on Facebook, Instagram, LinkedIn, Twitter, Twitch
