Rethinking your Kafka footprint: consolidation, migration, and cost control
Running multiple Kafka clusters gets expensive fast, and most teams end up with more infrastructure than they need. In this session, Karel Sague walks through a practical methodology for reducing Kafka costs by decommissioning old clusters, consolidating workloads, splitting clusters where it makes sense, and migrating to more cost-effective infrastructure.
Karel covers how to evaluate whether consolidation or splitting is the right call for a given workload, the sequencing and risk management involved in a live cluster migration, and how to avoid downtime and data loss along the way. The session includes code samples from real migrations.
If you are managing Kafka infrastructure and looking for ways to cut costs without cutting corners, this talk is for you.
Speaker
Karel Sague
Staff Software Engineer, Factor House
Karel Sague is a Staff Software Engineer at Factor House, where he works on real-time data streaming infrastructure spanning Kafka, Flink, and Iceberg. Before joining Factor House, he spent 8 years as a Software Engineer at Square (Block Inc.), and prior to that worked as a consultant at ThoughtWorks. Across his career he has built and operated large-scale data platforms in financial services and consulting, with deep experience in event-driven architecture, streaming ingestion, and open table formats.
