In flink 1.17.1, I got a "FlinkException: TaskExecutor akka... has no more allocated alots for job xxx" error when I run a batch job in k8s with Flink k8s operator. I submit a flinkdeployment job and job config is : tm.slot=15, adaptive.auto-parallelism.max=12 When job is running, I see some source parallelism is 28，is more than 12，and taskmanager pull up sequentially，But soon the job failed，and taskmanager is all ready。 I tried increase "jobmanager.adaptive-scheduler.resource-stabilization-timeout" to 60s, not work. Then I try increase "tm.slot" to 30, worked, But I can't determine exactly how many slots each taskmanager needs, so I don't think this approach will work Is the reason that the job starts too fast and the slot resource is not ready?