I’m using PyFlink (1.18v) on Kubernetes pods and want to ensure that checkpointing is enabled so that if a pod restarts, the process continues from the last checkpoint. I have configured checkpointing to write to an AWS S3 bucket, and it appears to be working. However, after killing a pod, the process does not resume from the last checkpoint, leading to missing data in the Kafka stream after the pod restarts.

I’m using PyFlink (1.18v) on Kubernetes pods and want to ensure that checkpointing is enabled so that if a pod restarts, the process continues from the last checkpoint. I have configured checkpointing to write to an AWS S3 bucket, and it appears to be working. However, after killing a pod, the process does not resume from the last checkpoint, leading to missing data in the Kafka stream after the pod restarts.