Kestra 1.0: Achieving 99% Faster Queries and Massive Scale With Xiaomi

Kestra 1.0 introduces significant performance and scalability enhancements, primarily through optimizations in MySQL queries, worker thread scaling, and Kafka Streams, resulting in faster execution times, reduced latency, and increased throughput for both open-source and Enterprise deployments. The collaboration with Xiaomi led to notable improvements in MySQL query execution, drastically reducing time from 30 seconds to a few hundred milliseconds by replacing deprecated query methods. The default number of worker threads has been increased, improving throughput and reducing latency, particularly noticeable at higher execution rates. Additionally, enhancements to the Kafka backend now allow the system to better utilize available resources, doubling the maximum supported throughput to 4000 executions per minute. These upgrades underscore Kestra’s commitment to continually improving system performance and scalability, with further developments anticipated.