Serverless computing abstracts away server management and scales applications based on exact demand, including scaling to zero to reduce costs.
Compare microservices for long-lived, developer-controlled processes against serverless for short-lived, cloud-managed, event-driven functions.
Knative enables serverless, container-based workloads on Kubernetes by using standard primitives to manage scaling, revisions, and traffic.
Quarkus overcomes traditional Java performance issues in serverless environments by offering sub-second startup times and a minimal memory footprint.
Use the code.quarkus.io starter to quickly generate a new project with dependencies for JAX-RS, Spring Web, or AWS Lambda.
A live demo shows a native-compiled Quarkus application starting in milliseconds and automatically scaling down to zero pods after a period of inactivity.
A side-by-side comparison demonstrates that a native Quarkus application starts in milliseconds while a similar Spring Boot application takes several seconds.
Knative's revision system allows for safe deployment strategies like blue-green by managing traffic splitting between different versions of a service.
Configure a Knative service to automatically scale up the number of pods based on a target number of concurrent requests.
Knative Eventing can consume messages from a Kafka topic and automatically trigger and scale services to process events asynchronously.
Use tools like Telepresence or adjust the scale-to-zero timeout to connect a local debugger to a running service in a Kubernetes cluster.
Jobs that call for the skills explored in this talk.