Databricks cost monitoring across every workspace you run.
LakeSentry monitors Databricks cost from one place. It ingests system tables and API metadata from every workspace into a normalized cost ledger, tracks spend by team, job, cluster, and SQL warehouse, and raises alerts that carry their own evidence. Daily refresh by default, hourly extraction when you need it.
Read-only by default. Write access only if you choose to execute actions.
What “We Monitor Costs” Usually Means
In most Databricks organizations, cost monitoring means a person: someone who owns a dashboard, checks it when they remember, and gets asked “why is the bill up” twice a quarter. The tooling underneath is per-workspace usage data, so that person is also the integration layer.
The native building blocks are real: system tables are detailed and reliable. What’s missing is the layer that watches them continuously, across workspaces, with a memory of what normal looks like.
- Each workspace reports separately; the account-wide view is a join someone maintains.
- Dashboards answer the question you ask; they don’t raise the one you didn’t.
- Jobs, warehouses, serving, and storage costs live in different views.
How LakeSentry Monitors Databricks Cost
One ledger, views for every question, and alerts judged against each workload’s own history.
One normalized ledger
Billing records and workload metadata from every workspace, ingested into a single cost ledger with a consistent shape. Jobs, all-purpose compute, SQL warehouses, model serving, and storage: comparable numbers in one place.
Break it down without SQL
Break spend down by job, warehouse, compute type, or team, and pivot between views to isolate a change, with no SQL required. When you want the raw trail, the system-table lineage is still underneath.
Alerts with a memory
Seven anomaly detectors watch the ledger against per-workload baselines, from cost spikes to budget risk. An alert arrives with its baseline and deviation, so monitoring starts the investigation instead of just announcing one.
Native Dashboards vs a Monitoring Platform
System tables stay the source of truth either way. What changes is who watches them.
| Native usage dashboards | LakeSentry | |
|---|---|---|
| Scope | Per workspace | Every workspace in one normalized ledger |
| Attribution | Tags, where present | Rules, mappings, and usage-based splits, with confidence tiers |
| Alerting | Fixed budget thresholds | Statistical baselines with evidence in every alert |
| Waste review | Manual queries, when someone has time | Detected patterns, ranked by observed cost |
| Getting started | SQL and dashboard building | A read-only service principal — minutes, not days |
Go Deeper
The mechanics live in the docs; the reasoning lives on the blog.
How LakeSentry works
Architecture overview: collectors, ingestion, attribution, insights.
What native Databricks cost tools show
What system tables and dashboards do well, and where the gap starts.
Databricks cost tools, compared
The category overview: native tooling, LakeSentry, and the alternatives.
Monitoring is the core of every tier: unlimited workspaces from Free, longer history and full insight detail as you move up.
Free
$0€0
Standard
$499€499/mo
Pro
$849€849/mo
No per-DBU tax. No per-workspace fees. Paid tiers billed annually.
Frequently Asked Questions
Which Databricks costs does LakeSentry monitor?
How fresh is the data?
What do we need to build to start monitoring?
How does multi-workspace monitoring work?
What does LakeSentry read from our environment?
See it in your own environment.
The free tier covers unlimited workspaces with three months of history. No card required.