improve log message at waiting for pods

perfmetrics/scripts/testing_on_gke/examples/run-gke-tests.sh

@@ -522,9 +522,18 @@ function waitTillAllPodsComplete() {
       printf "All pods completed.\n\n"
       break
     else
-      printf "There are still "${num_noncompleted_pods}" pod(s) incomplete (either still pending or running). So, sleeping for now... will check again in "${pod_wait_time_in_seconds}" seconds.\n\n"
+      printf "${num_noncompleted_pods} pod(s) is/are still pending or running. Will check again in "${pod_wait_time_in_seconds}" seconds. Sleeping for now.\n"
       printf "To ssh to any specific pod, use the following command: \n"
-      printf "  kubectl exec -it pods/<podname> --namespace=${appnamespace} -- /bin/bash \n\n"
+      printf "  gcloud container clusters get-credentials ${cluster_name} --location=${zone}\n"
+      printf "  kubectl config set-context --current --namespace=${appnamespace}\n"
+      printf "  kubectl exec -it pods/<podname> [-c {gke-gcsfuse-sidecar|fio-tester|dlio-tester}] --namespace=${appnamespace} -- /bin/bash \n"
+      printf "To view cpu/memory usage of different pods/containers: \n"
+      printf "  kubectl top pod [<podname>] --namespace=${appnamespace} [--containers] \n"
+      printf "To view the latest status of the pods: \n"
+      printf "  kubectl get pods --namespace=${appnamespace} -o wide [--watch] \n"
+      printf "To refer back to this run for fetching/parsing outputs, use environment variables "
+      printf " \" output_dir=${output_dir} cluster_name=${cluster_name} workload_config=${workload_config} instance_id=${instance_id} \""
+      printf " with ${0}\n\n\n\n"
     fi
     sleep ${pod_wait_time_in_seconds}
     unset podslist # necessary to update the value of podslist every iteration