Hello Flink Community, We run the Flink Kubernetes Operator 1.13.0 with Flink 2.2 (Application mode) and the built-in Autoscaler enabled on AWS EKS, using Karpenter for node provisioning and consolidation. We're hitting a collision between Karpenter's consolidation and the Flink Operator Autoscaler: when the Autoscaler rescales a job, nodes briefly appear underutilized, so Karpenter consolidates and evicts active TaskManager/JobManager pods. This leads to cluster instability and abnormal behavior that does not seem to recover without manual intervention. To mitigate the issue and avoid aggressive pod evictions (especially TaskManagers), we annotated the TM pods with karpenter.sh/do-not-disrupt: "true". However, this introduced a new problem: some nodes are no longer consolidated at all.
Questions: 1. What is the recommended pattern for running the Autoscaler alongside Karpenter? 2. JobManagers are currently configured with PodDisruptionBudgets. Should TaskManagers have PDBs as well? 3. Would you recommend isolating Flink onto a dedicated Karpenter NodePool with consolidationPolicy: WhenEmpty? We're currently considering do-not-disrupt annotations, PDBs, a dedicated NodePool , and a larger stabilization window. Your feedback on whether these align with community best practices would be very welcome. Thanks! Tamir Confidentiality: This communication and any attachments are intended for the above-named persons only and may be confidential and/or legally privileged. Any opinions expressed in this communication are not necessarily those of NICE Actimize. If this communication has come to you in error you must take no action based on it, nor must you copy or show it to anyone; please delete/destroy and inform the sender by e-mail immediately. Monitoring: NICE Actimize may monitor incoming and outgoing e-mails. Viruses: Although we have taken steps toward ensuring that this e-mail and attachments are free from any virus, we advise that in keeping with good computing practice the recipient should ensure they are actually virus free.
