Explanation
Why the waste happens and who it affects.
To mitigate idle charges, Databricks provides an "autotermination" setting that shuts down clusters after a period of inactivity. However, if the termination period is set too high, or if policies do not enforce reasonable thresholds, idle clusters can persist for long durations without performing any work - resulting in wasted compute spend. Lowering the termination window reduces exposure to idle time while preserving user flexibility.
Billing model
The pricing dimensions that drive this cost.
Databricks compute is billed via:
- Databricks Units (DBUs): Rated per hour based on compute type and node configuration, and billed per second
- Underlying Cloud Infrastructure: Charged based on VM runtime, billed per second or minute
Interactive clusters continue accruing DBU and VM costs while running, including during idle periods. The autotermination setting determines how long idle clusters persist before being shut down automatically. Longer termination windows result in more idle time and unnecessary cost.
How to detect
3 checks to find it in your estate.
- List all interactive clusters and review their configured autotermination settings
- Compare termination time thresholds to workspace policy guidelines or best practices
- Confirm whether any long windows are tied to justified use cases (e.g., infrequent but heavy dev work)
How to fix
3 ways to remove the waste.
- Lower the autotermination threshold for interactive clusters
- Apply workspace compute policies to cap the maximum idle time for clusters
- Grant exceptions only when use cases are documented and cost impact is understood
Documentation
Vendor references for pricing and configuration.