In this episode, we’re diving deep into Kubernetes autoscaling—but with a twist.
🚀 What You’ll Learn:
• The difference between HPA and VPA, and when to use each
• Why default CPU/memory metrics can be misleading
• How to use external metrics (like traffic volume) for more accurate scaling
• How to correlate resource usage with application health using OpenTelemetry and Dynatrace
• The magic of Pearson correlation to identify real scaling opportunities
• How to automate scaling decisions with DQL, EdgeConnect, and GitOps workflows
🔧 Tools & Concepts Covered:
• Kubernetes HPA & VPA
• Dynatrace Operator & DQL
• OpenTelemetry Collector
• KEDA & KEPN
• EdgeConnect for secure cluster interaction
• GitOps-friendly scaling policies
📈 Whether you're running microservices at scale or just getting started with Kubernetes, this episode will help you build a data-driven autoscaling strategy that actually works.
---------------------------
📖 Chapters 📖
00:00 Intro
00:50 Autoscaling 101
01:45 Default Scaling Metrics – Limitations
02:33 What If You Stick with Built-in Metrics?
02:47 Observing the Problem
03:22 Correlating Metrics – The Secret Sauce
03:36 The Pierson correlation number
04:05 DQL
04:30 Worflow
05:16 Edge Connect
05:39 Conclusion
-----------------------------
🔗 Additional Links
Dynatrace Trial
Tutorial
-----------------------------
Subscribe to our YT channel
Stay up-to-date with Dynatrace! Follow us on Facebook, Instagram, LinkedIn, Twitter, Twitch
