Conference talks and technical presentations on distributed systems, Kafka and Kubernetes.
What if losing an entire Kubernetes cluster didn't take your Kafka down with it? We built a working stretch Kafka cluster distributed across three independent Kubernetes clusters. In our full-cluster failure test, we terminated every pod in one cluster simultaneously — Kafka stayed operational, producers and consumers saw no interruption, and a new KRaft leader was elected in 1.5 seconds.
As organisations rely on Kafka for mission-critical workloads, ensuring high availability across Kubernetes clusters is a key challenge. This talk explores Stretch Clusters — a proposed Strimzi enhancement that distributes Kafka brokers across clusters while maintaining a unified deployment model, leveraging cloud-native networking solutions like Submariner and Cilium.
A Kubernetes cluster outage shouldn't mean a Kafka outage. This session explores practical methodologies for deploying Kafka across multiple Kubernetes clusters — covering broker and controller distribution, fault tolerance, cross-cluster communication, and resource synchronisation within the Operator framework.
I speak at conferences and meetups about Kafka, Kubernetes, and distributed systems. If you're organising an event and would like to discuss a talk topic, feel free to reach out.
Get in touch