All the recently terminated clusters are displayed in the ‘Terminated Clusters’ section in the cluster list page. The section displays clusters terminated in the last 7 days with a max cap of 70 such clusters. The bottom of the section also displays up to 30 jobs clusters that are recently terminated by jobs scheduler.
You can manually terminate a cluster both from the cluster list and details 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, then Databricks will automatically terminate that cluster.
A cluster is inactive if all commands on the cluster have finished executing, including Spark Jobs, Structured Streaming, and JDBC. This does not include commands run by SSH-ing into the cluster and running bash commands.
You may opt out of auto-termination by unselecting the checkbox above, or by specifying an inactivity period of “0” in the API.
Auto-termination is best supported in the latest Spark versions. Older spark versions have known limitations which may result in inaccurate reporting of cluster activity. For example, clusters running JDBC, R, or streaming commands may report a stale activity time which will lead to premature cluster termination. Users are strongly recommended to upgrade to the most recent spark version to benefit from bug fixes and improvements to the autotermination feature.
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. Please turn off autotermination for clusters running DStreams or consider using Structured Streaming.
Databricks records information whenever a cluster is terminated.
This page lists common termination reasons and describes potential steps for remediation.
Cloud Provider Launch Failures¶
This termination reason occurs when Databricks fails to acquire virtual machines from Azure. The error code and message from the API are propagated to help you troubleshoot the issue.
This means that Databricks was able to launch the cluster, but lost connection to the instance hosting the Spark driver.
This may be caused by the driver virtual machine going down, or a networking issue.