MetricSign
Start free
Medium severityclusterDatabricks

Databricks Error:
INSTANCE_POOL_MAX_CAPACITY_REACHED

What does this error mean?

A Databricks cluster could not be created or scaled because the instance pool it is attached to has reached its configured maximum capacity. No additional instances can be allocated from the pool.

Common causes

  • 1Multiple jobs started simultaneously, all requesting clusters from the same pool
  • 2The pool maximum capacity is set too low for peak concurrent workloads
  • 3A long-running job is holding pool instances that would otherwise be available for new clusters
  • 4Auto-scaling of existing clusters consumed all remaining pool capacity

How to fix it

  1. 1Open Databricks > Compute > Pools, select the pool, and increase the Maximum Capacity setting.
  2. 2Check which clusters are currently drawing from the pool: go to Compute > Clusters and filter by pool name. Terminate idle clusters to free capacity immediately.
  3. 3Stagger job start times in the job schedules to spread cluster creation across a wider window.
  4. 4For lower-priority jobs, create a separate pool with a lower max capacity so they don't compete with critical pipelines.
  5. 5As a longer-term fix, enable 'Fall back to on-demand instances' in the pool settings so jobs can still run when the pool is exhausted — at higher cost.

Frequently asked questions

Does the pool automatically increase capacity?

No — the pool maximum capacity is a hard limit set by an admin. It must be manually increased in the pool settings.

Can I monitor pool utilization?

Yes — the Databricks UI shows current pool utilization. You can also use the Instance Pools API to query pool state programmatically.

Source · docs.databricks.com/aws/en/clusters/cluster-error-codes.html

Other cluster errors