This article describes how to manage Databricks clusters, including displaying, editing, starting, terminating, deleting, controlling access, and monitoring performance and logs.
To display the clusters in your workspace, click Compute in the sidebar.
The Compute page displays clusters in two tabs: All-Purpose Clusters and Job Clusters.
For each cluster, this page shows:
- Cluster name
- Number of nodes
- Databricks Runtime version
- Type of driver and worker nodes
- Cluster creator or job owner
In addition to the common cluster information, the All-Purpose Clusters tab shows the numbers of notebooks attached to the cluster. Above the list is the number of pinned clusters.
- Starting , Terminating
- Standard cluster
- High concurrency cluster
- Access Denied
- Table ACLs enabled
You can filter the cluster lists using the buttons and Filter field at the top right:
- To display only clusters that you created, click Created by me.
- To display only clusters that are accessible to you (if cluster access control is enabled), click Accessible by me.
- To filter by a string that appears in any field, type the string in the Filter text box.
30 days after a cluster is terminated, it is permanently deleted. To keep an all-purpose cluster configuration even after a cluster has been terminated for more than 30 days, an administrator can pin the cluster. Up to 70 clusters can be pinned.
You can pin a cluster from the cluster list or the cluster detail page:
To pin or unpin a cluster, click the pin icon to the left of the cluster name.
To pin or unpin a cluster, click the pin icon to the right of the cluster name.
You can also invoke the Pin API endpoint to programmatically pin a cluster.
Sometimes it can be helpful to view your cluster configuration as JSON. This is especially useful when you want to create similar clusters using the Clusters API 2.0. When you view an existing cluster, simply go to the Configuration tab, click JSON in the top right of the tab, copy the JSON, and paste it into your API call. JSON view is ready-only.
You edit a cluster configuration from the cluster detail page. To display the cluster detail page, click the cluster name on the Compute page.
You can also invoke the Edit API endpoint to programmatically edit the cluster.
- Notebooks and jobs that were attached to the cluster remain attached after editing.
- Libraries installed on the cluster remain installed after editing.
- If you edit any attribute of a running cluster (except for the cluster size and permissions), you must restart it. This can disrupt users who are currently using the cluster.
- You can edit only running or terminated clusters. You can, however, update permissions for clusters that are not in those states on the cluster details page.
For detailed information about cluster configuration properties you can edit, see Configure clusters.
You can create a new cluster by cloning an existing cluster.
Cluster detail page
The cluster creation form is opened prepopulated with the cluster configuration. The following attributes from the existing cluster are not included in the clone:
- Cluster permissions
- Installed libraries
- Attached notebooks
Cluster access control within the Admin Console allows admins and delegated users to give fine-grained cluster access to other users. There are two types of cluster access control:
Cluster creation permission: Admins can choose which users are allowed to create clusters.
Cluster-level permissions: A user who has the Can manage permission for a cluster can configure whether other users can attach to, restart, resize, and manage that cluster by clicking the icon in the cluster actions.
To learn how to configure cluster access control and cluster-level permissions, see Cluster access control.
Apart from creating a new cluster, you can also start a previously terminated cluster. This lets you re-create a previously terminated cluster with its original configuration.
You can start a cluster from the cluster list, the cluster detail page, or a notebook.
Cluster detail page
Notebook cluster attach drop-down
You can also invoke the Start API endpoint to programmatically start a cluster.
Databricks identifies a cluster with a unique cluster ID. When you start a terminated cluster, Databricks re-creates the cluster with the same ID, automatically installs all the libraries, and re-attaches the notebooks.
When a job assigned to an existing terminated cluster is scheduled to run or you connect to a terminated cluster from a JDBC/ODBC interface, the cluster is automatically restarted. See Create a job and JDBC connect.
Cluster autostart allows you to configure clusters to autoterminate without requiring manual intervention to restart the clusters for scheduled jobs. Furthermore, you can schedule cluster initialization by scheduling a job to run on a terminated cluster.
If your cluster was created in Databricks platform version 2.70 or earlier, there is no autostart: jobs scheduled to run on terminated clusters will fail.
To save cluster resources, you can terminate a cluster. A terminated cluster cannot run notebooks or jobs, but its configuration is stored so that it can be reused (or—in the case of some types of jobs—autostarted) at a later time. You can manually terminate a cluster or configure the cluster to automatically terminate after a specified period of inactivity. Databricks records information whenever a cluster is terminated. When the number of terminated clusters exceeds 150, the oldest clusters are deleted.
Unless a cluster is pinned, 30 days after the cluster is terminated, it is automatically and permanently deleted.
When you run a job on a New Job Cluster (which is usually recommended), the cluster terminates and is unavailable for restarting when the job is complete. On the other hand, if you schedule a job to run on an Existing All-Purpose Cluster that has been terminated, that cluster will autostart.
You can manually terminate a cluster from the cluster list or the cluster detail page.
Cluster detail page
You can also set auto termination for a cluster. During cluster creation, you can specify an inactivity period in minutes after which you want the cluster to terminate. If the difference between the current time and the last command run on the cluster is more than the inactivity period specified, Databricks automatically terminates that cluster.
A cluster is considered inactive when all commands on the cluster, including Spark jobs, Structured Streaming, and JDBC calls, have finished executing. This does not include commands run by SSH-ing into the cluster and running bash commands.
- Clusters do not report activity resulting from the use of DStreams. This means that an autoterminating cluster may be terminated while it is running DStreams. Turn off auto termination for clusters running DStreams or consider using Structured Streaming.
- The auto termination feature monitors only Spark jobs, not user-defined local processes. Therefore, if all Spark jobs have completed, a cluster may be terminated even if local processes are running.
- Idle clusters continue to accumulate DBU and cloud instance charges during the inactivity period before termination.
You configure automatic termination in the Auto Termination field in the Autopilot Options box on the cluster creation page:
The default value of the auto terminate setting depends on whether you choose to create a standard or high concurrency cluster:
- Standard clusters are configured to terminate automatically after 120 minutes.
- High concurrency clusters are configured to not terminate automatically.
You can opt out of auto termination by clearing the Auto Termination checkbox or by specifying an inactivity period of
Auto termination is best supported in the latest Spark versions. Older Spark versions have known limitations which can result in inaccurate reporting of cluster activity. For example, clusters running JDBC, R, or streaming commands can report a stale activity time that leads to premature cluster termination. Please upgrade to the most recent Spark version to benefit from bug fixes and improvements to auto termination.
Sometimes a cluster is terminated unexpectedly, not as a result of a manual termination or a configured automatic termination.
For a list of termination reasons and remediation steps, see the Knowledge Base.
Deleting a cluster terminates the cluster and removes its configuration.
You cannot undo this action.
You cannot delete a pinned cluster. In order to delete a pinned cluster, it must first be unpinned by an administrator.
To delete a cluster, click the icon in the cluster actions on the Job Clusters or All-Purpose Clusters tab.
You can also invoke the Permanent delete API endpoint to programmatically delete a cluster.
Detailed information about Spark jobs is displayed in the Spark UI, which you can access from:
- The cluster list: click the Spark UI link on the cluster row.
- The cluster details page: click the Spark UI tab.
You can get details about active and terminated clusters.
If you restart a terminated cluster, the Spark UI displays information for the restarted cluster, not the historical information for the terminated cluster.
Databricks provides three kinds of logging of cluster-related activity:
- Cluster event logs, which capture cluster lifecycle events, like creation, termination, configuration edits, and so on.
- Apache Spark driver and worker logs, which you can use for debugging.
- Cluster init-script logs, valuable for debugging init scripts.
This section discusses cluster event logs and driver and worker logs. For details about init-script logs, see Init script logs.
The cluster event log displays important cluster lifecycle events that are triggered manually by user actions or automatically by Databricks. Such events affect the operation of a cluster as a whole and the jobs running in the cluster.
For supported event types, see the REST API ClusterEventType data structure.
Events are stored for 60 days, which is comparable to other data retention times in Databricks.
Click Compute in the sidebar.
Click a cluster name.
Click the Event Log tab.
To filter the events, click the in the Filter by Event Type… field and select one or more event type checkboxes.
Use Select all to make it easier to filter by excluding particular event types.
The direct print and log statements from your notebooks, jobs, and libraries go to the Spark driver logs. These logs have three outputs:
- Standard output
- Standard error
- Log4j logs
To access these driver log files from the UI, go to the Driver Logs tab on the cluster details page.
Log files are rotated periodically. Older log files appear at the top of the page, listed with timestamp information. You can download any of the logs for troubleshooting.
To view Spark worker logs, you can use the Spark UI. You can also configure a log delivery location for the cluster. Both worker and cluster logs are delivered to the location you specify.
To help you monitor the performance of Databricks clusters, Databricks provides access to Ganglia metrics from the cluster details page.
You can also install Datadog agents on cluster nodes to send Datadog metrics to your Datadog account.
To access the Ganglia UI, navigate to the Metrics tab on the cluster details page. CPU metrics are available in the Ganglia UI for all Databricks runtimes. GPU metrics are available for GPU-enabled clusters.
To view live metrics, click the Ganglia UI link.
To view historical metrics, click a snapshot file. The snapshot contains aggregated metrics for the hour preceding the selected time.
You can install Datadog agents on cluster nodes to send Datadog metrics to your Datadog account. The following notebook demonstrates how to install a Datadog agent on a cluster using a cluster-scoped init script.
To install the Datadog agent on all clusters, use a global init script after testing the cluster-scoped init script.
This feature is available on Databricks Runtime 8.0 and above.
Because spot instances can reduce costs, creating clusters using spot instances rather than on-demand instances is a common way to run jobs. However, spot instances can be preempted by cloud provider scheduling mechanisms. Preemption of spot instances can cause issues with jobs that are running, including:
- Shuffle fetch failures
- Shuffle data loss
- RDD data loss
- Job failures
You can enable decommissioning to help address these issues. Decommissioning takes advantage of the notification that the cloud provider usually sends before a spot instance is decommissioned. When a spot instance containing an executor receives a preemption notification, the decommissioning process will attempt to migrate shuffle and RDD data to healthy executors. The duration before the final preemption is typically 30 seconds to 2 minutes, depending on the cloud provider.
Databricks recommends enabling data migration when decommissioning is also enabled. Generally, the possibility of errors decreases as more data is migrated, including shuffle fetching failures, shuffle data loss, and RDD data loss. Data migration can also lead to less re-computation and save cost.
Decommissioning is best effort and does not guarantee that all data can be migrated before final preemption. Decommissioning cannot guarantee against shuffle fetch failures when running tasks are fetching shuffle data from the executor.
With decommissioning enabled, task failures caused by spot instance preemption are not added to the total number of failed attempts. Task failures caused by preemption are not counted as failed attempts because the cause of the failure is external to the task and will not result in job failure.
To enable decommissioning, you set Spark configuration settings and environment variables when you create a cluster:
To enable decommissioning for applications:
To enable shuffle data migration during decommissioning:
spark.storage.decommission.enabled true spark.storage.decommission.shuffleBlocks.enabled true
To enable RDD cache data migration during decommissioning:
When RDD StorageLevel replication is set to more than 1, Databricks does not recommend enabling RDD data migration since the replicas ensure RDDs will not lose data.
spark.storage.decommission.enabled true spark.storage.decommission.rddBlocks.enabled true
To enable decommissioning for workers:
To set these custom Spark configuration properties:
- On the New Cluster page, click the Advanced Options toggle.
- Click the Spark tab.
To access a worker’s decommission status from the UI, navigate to the Spark Cluster UI - Master tab:
When the decommissioning finishes, the executor that decommissioned shows the loss reason in the Spark UI > Executors tab on the cluster’s details page: