Access Mosaic Sentinel Data Using a Direct JDBC Connection

Starting in Strategy (September 2026), you can access Mosaic Sentinel data using a direct JDBC connection, allowing you to seamlessly integrate the data into your own applications using SQL.

Prerequisites

You must have the Use Mosaic Sentinel and Monitor Activity privileges.

Explore Data

You can use any SQL database tool (for example, DBeaver) to explore the Mosaic Sentinel catalog by establishing a JDBC connection to your semantic layer.

Connect to a JDBC driver. For more information, see Connect to a JDBC Driver.
Use SQL queries from the tables under the sentinel.views schema.

Each sentinel table includes a tenant_id column to support filter logic on the backend. You can disregard this column.

Data Dictionary

See the following data your can view in your Mosaic Sentinel tables:

sentinel_audit (Audit & Compliance)

Data	Column Name	Data Type	Description
Timestamp (UTC)	`action_dt`	varchar	When the action occurred in UTC.
Actor	`actor`	varchar	Name of the user who performs the action.
Action	`action_type`	varchar	What action was completed (for example, EXECUTE, SAVE, DELETE).
Object Name	`object_name`	varchar	Name of the object that was used.
Object Type	`object_type`	varchar	Type of object (for example, Metric, Attribute, Mosaic model, Security Filter, or raw type).
Sensitive Data	`sensitive_data`	varchar	Whether the object has sensitive data, personal information, or none.
Object ID	`object_guid`	varchar	GUID of the object.
Model Name	`model_name`	varchar	Related model or cube name, if any.
Model Type	`model_type`	varchar	Type of model (for example, Intelligent Cube, Live Intelligent Cube, Mosaic model, Data Import, or raw type).
Certified Model	`is_certified`	varchar	Whether the related model is certified (for example, Certified or Not Certified).
Project	`project_name`	varchar	Project the object belongs to.
Status	`status`	varchar	Action outcome (for example, Success, Failed, Unknown, or a status message).

sentinel_risk (Risk Management)

Data	Column Name	Data Type	Description
Timestamp (UTC)	`action_dt`	varchar	When the anomalous access occurred in UTC.
Anomaly Type	`anomaly_type`	varchar	Type of anomaly (for example, unusual time, unusual frequency, sensitive data, or large-volume sensitive access).
Data Accessed	`data_accessed`	varchar	Object names that were accessed in the event.
User	`user_name`	varchar	Name of the user who triggered the alert.
User ID	`user_guide`	varchar	GUID of the user.
User Group	`user_groups`	varchar	Groups the user belongs to.
Application	`application_type`	varchar	Application the access came from.
Mosaic Model	`model_name`	varchar	Name of the Mosaic model the object was accessed in.
Mosaic Model ID	`model_guid`	varchar	GUID of the Mosaic model.
Description	`description`	varchar	Description of the alert.

sentinel_usage_insights_model (Usage Insights - Model)

Data	Column Name	Data Type	Description
Model Name	`model_guid`	varchar	Name of the model or cube.
Model Type	`model_type`	varchar	Type of model (for example, Intelligent Cube, Live Intelligent Cube, Mosaic Model, Data Import, or raw type).
Owner	`owner_name`	varchar	Display name of the model owner.
Sources	`first_src_name`	varchar	Data source in the model (first source and the number of additional sources).
Number of Executions	`executions`	bigint	The number of times the model was executed (excluding linked sessions).
Number of Users	`users`	bigint	The number of distinct users executed in the model.

sentinel_usage_insights_source (Usage Insights - Source)

Data	Column Name	Data Type	Description
Source Name	`src_name`	varchar	Name of the data source.
Source Type	`src_type`	varchar	Type of source (for example, Snowflake, Databricks, PostgreSQL, etc).
Number of Execution	`COUNT(model_guid)`	bigint	The number of executions that hit the source (excluding linked sessions).
Number of Data Models	`COUNT(DISTINCT model_guid)`	bigint	The number of distinct models that use the source.
Number of Users	`COUNT(DISTINCT actor_guid)`	bigint	The number of distinct users that are executed against the source.

sentinel_usage_insights_object (Usage Insights - Object)

Data	Column Name	Data Type	Description
Object Name	`object_name`	varchar	Name of the object.
Object Type	`object_type`	varchar	Type of object (Metric, Attribute, Mosaic Model, Security Filter).
Model Name	`model_name`	varchar	Related model or cube name.
Owner	`owner_name`	varchar	Display name of the object owner.
Number of Executions	`COUNT(DISTINCT log_id)`	bigint	The number of times the object was executed (excluding linked sessions).
Number of Users	`COUNT(DISTINCT actor_guid)`	bigint	The number of distinct users executed in the object.

sentinel_cost (Cost Intelligence)

Data	Column Name	Data Type	Description
Model	`COALESCE(object_name, dataset_id)`	varchar	Model display name.
Query Mode	`serve_mode`	varchar	Whether the model ran as Import or Live.
Model Size	`model_size_bytes`	bigint	Model size in bytes.
Query Source	`datasource_type`	varchar	Data source type used by the model.
In-Memory Executions	`in_memory_executions`	bigint	Cube or in-memory executions over the last 30 days.
Source Executions	`source_executions`	bigint	Live executions over the last 30 days.
Execution Trend	`execution_trend`	array(varchar)	Daily live-execution counts over the last 30 days (for a sparkline).
Cost Trend	`cost_trend`	array(varchar)	Daily estimated cost over the last 30 days (for a sparkline).
Estimated Cost	`est_cost`	double	Total estimated cost over the last 30 days.
Potential Savings	`savings_estimate`	double	Latest savings estimate for this model, if any.