Event-Driven Autoscaling for Serverless Java on Kubernetes

Explore event-driven autoscaling for serverless Java applications in this 48-minute conference talk from Spring I/O 2024. Dive into the limitations of traditional Kubernetes autoscaling based on hardware resource utilization and discover how KEDA and Knative can address these challenges. Learn to redesign your Kubernetes autoscaling architecture using event-driven metrics from Apache Kafka, moving beyond standard CPU and memory resources. Gain insights into integrating Knative and KEDA for more efficient serverless Java deployments, enabling autoscaling based on event metrics from multiple sources like Apache Kafka and AWS SQS. Understand how this approach can enhance reliability and stability across various business use cases, including web apps, mobile applications, IoT edge streaming, and AI/ML workloads.

Event-driven autoscaling for Serverless Java by Daniel Oh @ Spring I/O 2024