Run LakeSentry inside your own Databricks account.
Deploy LakeSentry as a native Databricks App, now available on the Databricks Marketplace. The web app, cost ledger, insights engine, and MCP server all run on your Databricks runtime. Usage data, query text, and credentials never leave your environment.
Selected as one of 15 third-party apps in the Databricks Marketplace Apps launch at Data + AI Summit 2026.
Install from the Databricks Marketplace, with native identity, networking, and governance.
Prefer a hosted setup? Start with the SaaS.
- Web App
- Cost Ledger
- Insights
- MCP Server
Why teams pick the Databricks App.
The same LakeSentry product, deployed for organizations that need data to never leave their account.
Data sovereignty by default
Cost ledger, insights, and query text live in your Lakebase. The only outbound call is a periodic license check.
Native identity & governance
Authenticate through your existing Databricks SSO. Inherit Unity Catalog permissions, with no extra IAM to manage.
Vetted by Databricks
Databricks security-reviewed LakeSentry before publishing it to the Marketplace. That is one less vendor binary for your team to evaluate.
No connector to maintain
You skip the service principal, the read-only role, and the IP allowlist entirely. The app reads system tables directly, in-region.
Four components. One Databricks App.
The same architecture as our SaaS, minus the hosting boundary. Everything materializes inside Unity Catalog and is governed by your existing policies.
-
Web App
The full LakeSentry UI for dashboards, insights, approvals, and audit, served from your Databricks workspace.
-
Cost Ledger
Normalized job, pipeline, SQL, and compute usage materialized in Lakebase tables you own and govern.
-
Insights engine
Anomaly detection, waste identification, and significant-work ranking, computed in-account on a schedule.
-
MCP Server
Expose LakeSentry as a tool surface to Databricks Genie or your own AI agents, with full Unity Catalog scoping.
Get started.
Two ways to deploy LakeSentry into your Databricks account — pick the one that fits your workflow.
- 1
Provision a Lakebase instance
LakeSentry materializes its cost ledger in your own Lakebase database, which is PostgreSQL-compatible and fully under your governance. Spin one up in your workspace.
- 2
Create an encryption-key secret
Store a personal encryption key in a Databricks secret. LakeSentry uses it to encrypt sensitive data at rest, inside your account.
databricks secrets put-secret <SCOPE> <SECRET> --string-value "$(openssl rand -base64 24)" - 3
Install from the Databricks Marketplace
Install the LakeSentry app and point it at the secret name and Lakebase instance from the first two steps. Nothing leaves your account.
Always at parity with the SaaS version, the script installer gives you full control over how LakeSentry is installed and managed in your Databricks account — with no Marketplace dependency.
- 1
Prerequisites
Before running the installer, make sure you have the following tools installed:
- databricks CLI — installed and configured with a profile (install guide)
- curl
- openssl
- jq or python3 — at least one is required
Your Databricks account also needs permission to manage Apps, create secret scopes and secrets, and write to workspace files.
- 2
Install
Run this command in your terminal:
bash <(curl -fsSL https://github.com/lakesentry/lakesentry-releases/releases/latest/download/install.sh)The installer will ask two questions:
- Databricks CLI profile — press Enter to use DEFAULT, or enter the name of your configured profile.
- App name — press Enter to accept lakesentry, or choose a custom name.
Everything else is automatic: the app binary is downloaded and uploaded to your workspace, an encryption key is generated and stored in Databricks Secrets, and the Databricks App is created.
- 3
Save your encryption key
When the installer finishes it prints your encryption key. Copy it and store it somewhere safe. You will need it if you ever reinstall and need to recover access to existing data.
- 4
Provision Lakebase database
Before the app can start, attach a Lakebase database resource in your Databricks workspace:
- 1 Go to Compute → Apps → [your app name] → Edit
- 2 Under Resources, click Add Resource:
- Type: Lakebase Database (Autoscaling)
- Key: database
- 3 Click Save, then deploy the app from the UI or run the deploy command the installer printed.
- 1
- 5
Start the app
The installer printed a deploy command at the end — copy it from your terminal output and run it:
databricks app deploy <APP_NAME> --source-code-path /Workspace/...
Advanced options
Prefix the install command with any of these environment variables to change behavior:
| Variable | Description |
|---|---|
| BUNDLE_URL=... | Install a specific version instead of the latest. |
| DBR_PROFILE=... | Use a specific Databricks CLI profile. |
| APP_NAME=... | Use a custom app name. |
| ENCRYPTION_KEY=... | Restore a known key — for recovery or reinstall with existing data. |
| ROTATE_KEY=true | Generate a new key — ⚠ makes existing encrypted data unreadable. |
Recovery example — if you need to reinstall and restore access to existing data, supply your saved key:
ENCRYPTION_KEY=your64hexkeyhere \
bash <(curl -fsSL https://github.com/lakesentry/lakesentry-releases/releases/latest/download/install.sh)The installer is safe to re-run at any time. Existing encryption keys and app configuration are preserved by default.
Databricks App FAQ
The questions teams ask before installing LakeSentry from the Marketplace.
Do I need Unity Catalog?
Yes. LakeSentry reads cost and usage data from the Unity Catalog system tables, so your workspace needs it enabled. The app then authenticates through your existing Databricks SSO.
Does my data ever leave my account?
No. The web app, cost ledger, insights engine, and MCP server all run on your Databricks Apps runtime, and the cost ledger lives in your own Lakebase database. The only outbound call is a periodic license check.
How do I install the Databricks App?
Provision a Lakebase instance, create a Databricks secret holding a personal encryption key, then install LakeSentry from the Databricks Marketplace and provide the secret name and Lakebase instance from the first two steps.
What database does it use?
LakeSentry's backend runs on PostgreSQL, so it integrates natively with Databricks Lakebase. Your normalized cost ledger materializes in Lakebase tables you own and govern under Unity Catalog.
How is it different from the SaaS?
The product is identical. With SaaS, LakeSentry is hosted and connects through a read-only service principal, so system-table snapshots leave your environment. With the Databricks App, everything runs in your account and nothing but a license heartbeat leaves it. It is the right choice when data sovereignty, procurement, or governance demand it.
Bring LakeSentry into your own Databricks account.
Install it from the Databricks Marketplace. Every component runs in your workspace, and your data never leaves it.