Hi guys, We enabled Flink autoscaler memory tuning but aren't seeing recommendations, even with clear over-provisioning. Versions: • Flink: v1_17 (1.17.1) • Operator: 1.13.0 Configuration:

job.autoscaler.enabled: "true"
job.autoscaler.memory.tuning.enabled: "true"  #also tried false for dry run
job.autoscaler.metrics.window: "5m"
job.autoscaler.stabilization.interval: "30s"
job.autoscaler.scaling.enabled: "false"

What we've done: 1. Removed explicit memory configs (taskmanager.memory.process.size, taskmanager.memory.managed.size, taskmanager.memory.network.max) 2. Set taskManager.resource.memory: 12g 3. Verified metrics collection (6+ data points in ConfigMap autoscaler-<deployment-name>) 4. Confirmed autoscaler is active (ScalingExecutor running, metrics being collected) Current situation: • Memory: 12GB allocated, ~24% heap usage (~1.36GB used) • Metrics: Being collected and stored in ConfigMap • No recommendations: Not appearing in operator logs, Kubernetes events, or FlinkDeployment status/annotations What we've checked: • Operator logs show autoscaler activity but no memory tuning recommendations • No errors or warnings related to memory tuning • Metrics window is filling (5+ minutes of data collected) • Job is processing data (not idle) Has anyone seen memory tuning recommendations with this operator version? Are there additional requirements or known issues? Any guidance on where recommendations should appear or how to troubleshoot?