You have a GKE cluster that has three TPU nodes. You are running an ML training job on the cluster, and you observe log entries in Cloud Logging similar to the one below. To identify the root cause of a performance issue, you need to view all stdout logs from the containers running on the TPU nodes. What should you do?
