Databricks compute, tied to the job that spent it.
A Databricks connector is on the CloudMonitor roadmap. It will read your usage into Fabric and attribute clusters, jobs, and storage to the team and workload behind them, with right-sizing and anomaly detection on top.
Databricks cost concentrates in a handful of clusters and jobs, and it moves fast when an interactive cluster runs idle or a job pattern changes. The Databricks integration will attribute that consumption to the team and workload that drove it, so the lakehouse bill becomes a set of decisions rather than a lump sum.
- Will attribute cluster, job, and SQL warehouse cost to teams and workloads.
- Will flag idle interactive clusters and oversized configurations.
- Will forecast daily spend and raise anomalies when a job pattern shifts.
- Will report alongside your Azure spend in one allocation model.
Workload attribution
Tie consumption to the job behind it.
The Databricks connector will map usage to the teams, jobs, and workloads that generated it, so a spike in spend points to a specific pipeline rather than a single platform total nobody owns.
- Cluster, job, and SQL warehouse cost attributed
- Team and workload ownership, not a platform lump sum
- Reconciled against the rest of your Azure estate
Right-sizing
Find the idle clusters and oversized jobs.
Interactive clusters left running and over-provisioned job configurations are the usual sources of lakehouse waste. CloudMonitor will surface them with the team attached, so the fix has an owner.
- Idle interactive clusters flagged for shutdown
- Oversized configurations surfaced with the owner
- Ranked by saving, the largest opportunity first
Anomaly detection
Catch a runaway job the day it changes.
A daily forecast per workload means a job that suddenly costs more, from a data-volume jump or a config change, will show up immediately, routed to the team behind it with the dollar impact attached.
- Per-workload daily forecast built on history
- Severity-tiered alerts tuned per team
- Routed to Teams, Jira, or ServiceNow
Databricks is one of the easiest places for cost to hide. A few clusters and jobs drive most of the bill, an idle interactive cluster burns money quietly, and platform spend rarely maps cleanly to the teams using it. The Databricks integration will read usage into your Fabric workspace and attribute it to the workloads behind it.
This connector is in planning. Because the lakehouse already lives close to Microsoft Fabric, that data will be read in place and modeled alongside your Azure spend. If Databricks coverage matters to your rollout, the beta team would like to know which workloads and SKUs you most need attributed.
More data platforms
Other integrations like this.
Be first to connect Databricks.
Join the beta and tell the team what you need from this connector.
