dataproc/Check Autoscaling Policy
Verify autoscaling policies.
All steps available in dataproc
Verify autoscaling policies.
Check for issues related to BigQuery connector such as version dependency conflicts.
Verify that the nodes in the cluster can communicate with each other.
Verify that the nodes in the cluster can communicate with each other.
Verify if the Dataproc cluster has quota issues.
Verify if Dataproc cluster has stockout issue.
Verify if the cluster version is supported.
Verify if STW GC Pause has happened.
Check for non-default GCS connector and for errors in logs connected to Cloud Storage.
Verify if dataproc job failed.
Verify if dataproc cluster init script failed.
Verify the presence of Job Throttling logs…
Verify if the killing of Orphaned applications has happened.
Checks if specified logs messages exist in the Dataproc cluster.
Check if OOM has happened on master.
Check if the permissions are set correctly.
Verify if the port exhaustion has happened.
Verify preemptibility.
Check if the subnetwork of the cluster has private google access enabled.
Check if the python import failure has happened.
Verify if dataproc cluster is using Shared VPC.
Check for logs indicating shuffle failures.
Verify the presence of shuffle service kill related logs.
Check if Stackdriver is enabled for the cluster.
Verify if secondary worker preemption has happened.
Verify if dataproc job failed due to task not found.
Verify if worker disk usage issue has happened.
Verify if OOM has happened on worker nodes.
Verify presence of CheckYarnRuntimeException logs…
The end step of the runbook
Gathers cluster parameters needed for further investigation.
Prepares the parameters required for the dataproc/cluster-creation runbook.
Verifies if the cluster is in Error state and gathers additional parameters.
Verify if cluster exists in Dataproc UI.
Check if the cluster is using internal IP only.
Prepares the parameters required for the dataproc/spark_job_failures runbook.
Validating service account and permissions in Dataproc cluster project or another project.
The end step of the runbook.