Hi all, We are facing a situation where Cruise Control’s memory usage keeps increasing over time, even though the actual cluster load has not changed. Our Kafka cluster is relatively small (around 230 partitions across 3 brokers), yet Cruise Control appears to consume a large amount of heap during proposal generation. While investigating, we noticed that our Cruise Control setup includes 17 goals (both hard and soft). From what we understand, simulating and evaluating all of these goals, especially the soft balancing goals, can introduce significant memory and CPU overhead. Before making any changes, I wanted to check with the community: Question: Do you commonly remove or disable some of Cruise Control’s soft goals in production to reduce the Analyzer’s computational load? If yes, which goals are generally considered reasonable or safe to deactivate without causing notable negative impact on cluster balance or operational behavior? Any real‑world experience or recommendations would be very helpful. Thanks!
