Home

Cost Chargeback (Databricks)

Note

This tab is visible only for Databricks clusters.

From this tab, you can create and view the Chargeback reports for the following:

  • Databricks Clusters (default)

    This option shows the chargeback cost attributed to cluster creators for the specified time range and filters organized based on the selected Group by option.

  • Job runs

    This option shows the chargeback cost attributed to job run creators for the specified period and filters organized based on the selected Group by option.

Generate chargeback report for Databricks clusters
cost-chargeback-cluster-donut-chart.png

Chargeback report for clusters

cost-chargeback-job-run-donut-chart.png

Chargeback report for job runs

  1. From the Chargeback Type dropdown, on the upper-right, select either Cluster or JobRun.

  2. Click schedule.png and select a time range to view the chargeback reports. You can choose any of the preset time ranges (Last 7 days, Last 14 days, Last 30 days) or select a custom time range, which can be between 1 day (min) and 90 days (max)

  3. From the Filter By dropdown, select any of the following options:

    • Workspace (default)

    • User

    • Clusters

    • Tags

      A tag comprises a key and a value. You must specify both the key and the value for each tag.

  4. You can drill down to a specific filter item from the adjacent dropdown to view the trend graph. For example, if you have selected the User filter earlier, you can choose user1, user2, user3, etc. You can select a maximum of up to 5 such filter values.

  5. Select any of the following Group By options. A Group by determines the category against which the total Chargeback cost of a selected Filter By option is calculated and organized.

    • User

    • Workspace

    • Cluster

    • Tag key

    For example, when you select a Filter by option as Workspaces and the group by option as Users, then the chargeback is computed for only the selected workspaces and then grouped by users.

    The details of the Cluster chargeback report are displayed in the following formats:

    • Donut charts: The donut charts show the chargeback results for the DBUs, Clusters/JobRuns, and Cost.

      Hover over a donut section to see the slice name, the value, and the percentage in the whole.

      Note

      There are times when the selected option cannot group a cluster/job run. In such cases, they are listed as NULL.

    • Tables: The chargeback report details are displayed in a table based on the selected Group By option.

      Columns

      Description

      User/Workspace/Cluster/Tag key

      The selected Group by option.

      DBUs

      Total DBU consumed by the selected Group by option, in units and percentage, for the corresponding Filter by option.

      Cost

      Cost calculated based on the estimated run durations of the underlying instances, DBUs, and VM price components for the selected Group By option. Storage costs are not included. The actual cost may differ.

      The cost units and percentage are shown.

      Cluster count

      The total active cluster count/job runs for the selected Group by option in units and percentage.

      Click optimize-button.png for the corresponding Group By option and you can view the details of all the jobs in the selected period. You can also click copy-url.png to copy the link to the Jobs page and view the Jobs for the specified period.

    Notice

    For the cost, you can filter by an attribute. For each of such attributes, you can specify multiple values. Unravel evaluates these values using the OR semantics. For example, if you filter by user attribute, you may select user A and user B. The results shown belong either to User A or to User B.

    For jobs, this behavior is different. You can filter by multiple attributes at the same time. For example, you may specify a workspace attribute and later a user attribute. Unravel evaluates those attributes using the AND semantics. However, within each attribute, you can specify multiple values. For instance, you may select User A and User B for the user attribute. Those are values evaluated using the OR semantics.