At a glance
Active SQL Sessions is the live count of open connections currently held against your Databricks SQL warehouses: BI tools, JDBC/ODBC clients, notebooks, and scheduled queries that have an established session. It is the concurrency dimension of warehouse load, sitting alongside saturation and query rate. Read it to understand how many distinct consumers are leaning on your SQL compute right now, and whether a session leak or a runaway BI refresh is quietly stacking up connections.
What it tracks
The card reports Active SQL Sessions for the selected period, sampled in real time. Each session is an established connection to a SQL warehouse from a client such as a dashboard tool, a JDBC/ODBC driver, a notebook, or a scheduled query. The figure is the count of sessions in an open state across all running warehouses at the moment of the read, refreshed continuously rather than averaged over a window. A steadily climbing count with no matching rise in query throughput usually means sessions are being opened and not closed (a connection leak in a client or pool), which exhausts warehouse concurrency slots even when little work is actually running. Read it against SQL Queries per Hour (live) to separate genuine demand from idle-session build-up, and against SQL Warehouse Saturation % to see whether those sessions are translating into queued work.Reconciling against the source
Cross-check in the Databricks workspace under SQL → SQL Warehouses → (select warehouse) → Monitoring, where the native connection and query history view lists active sessions, or query thesystem.query.history system table for the open-session series. Confirm you are reading the same warehouse set and time zone (Vortex IQ stores UTC and renders in your profile time zone) before treating any gap as real.