At a glance
The count of Databricks SQL warehouses currently in aRUNNING(orSTARTING) state in the connected workspace. SQL warehouses are the compute behind dashboards, BI tools, and ad-hoc SQL, and they bill on the SQL DBU SKU, separate from clusters. This is the fast answer to “how much query compute is live and billing right now?”
What it tracks
The value is a live count of SQL warehouses returned by the SQL Warehouses API where thestate is RUNNING or STARTING, polled in real time (RT) every 60 seconds. It deliberately excludes STOPPED, STOPPING, and DELETED warehouses, so the number reflects compute that is actually available to serve queries and consuming SQL DBUs.
Each warehouse counts as 1 regardless of its size (2X-Small through 4X-Large) or whether multi-cluster scaling has spun up extra clusters underneath it. That makes this a “how many query endpoints are alive” pulse rather than a cost figure. To weight it by spend, read it alongside DBU Burned (24h), and to see whether the live warehouses are actually busy, pair it with SQL Warehouse Saturation % and Active SQL Sessions. The classic signal here is a warehouse left running out of hours with auto-stop misconfigured, which is pure wasted DBU. Compute clusters are counted separately on Active Clusters.
Reconciling against the source
Confirm the live count in the workspace under SQL → SQL Warehouses (filter to Running), viadatabricks warehouses list on the Databricks CLI, or against the system.compute.warehouse_events system table for the historical record. A brief difference versus the UI is expected because Vortex IQ polls every 60 seconds.