Kubernetes Autoscaling: From Resources to Events, HTTP, and GPUs

Autoscaling is at the heart of running efficient, resilient, and cost-effective workloads on Kubernetes, yet getting it right is often harder than it seems. In this session, we’ll explore the spectrum of autoscaling approaches, from resource-driven (CPU/memory) to event- and HTTP-driven scaling, and even GPU-aware scenarios.

We’ll dive into what’s new in KEDA, how it extends Kubernetes autoscaling beyond traditional metrics, and share practical best practices drawn from real-world experience.

What You’ll Learn

• Resource-driven autoscaling: Traditional CPU and memory-based scaling patterns and their limitations
• Event-driven scaling: Leveraging external metrics and events for more intelligent scaling decisions
• HTTP-based scaling: Scaling based on HTTP request patterns and traffic metrics
• GPU-aware autoscaling: Special considerations and approaches for GPU workloads
• KEDA innovations: Latest features and capabilities in the KEDA ecosystem
• Best practices: Real-world insights and proven patterns for production deployments

Join us to discover how modern autoscaling approaches can transform your Kubernetes workloads, helping you achieve better performance, cost optimization, and operational efficiency.

Reduce cloud costs and complexity.

Start autoscaling with KEDA today.

Get Started