Skip to content

Click in-app to access the full platform documentation for your version of DataRobot.

User Activity Monitor reference

The following sections describe the fields returned for the User Activity Monitor (UAM), based on the selected report view. See the User Activity Monitor overview for information on using the tool.

Each row of both the online preview and download of reports relates to a single audited event. Data is presented in ascending order, with the earliest data appearing at the top of the report. You can download the reports using Export CSV. When exporting reports, you are prompted to filter records for report download. The filters you apply when previewing the report apply only to the online preview.

Hide sensitive information

When viewing App Usage or Prediction Usage reports, you can hide or display identifying information with the "Include identifying fields" option. You may want to hide the information, for example, if Customer Support will be accessing a report. If unchecked, the columns display in the report without values. The fields considered sensitive are marked with an asterisk in the tables below.

Admin Usage activity report

The Admin Usage activity output reports the following data about administrator operations and activities.

Report field Description
Timestamp—UTC Timestamp (UTC time standard) when the administrator event occurred
Event Type of administrator event, such as Create Account, Organization created, Change Password, Update Account, etc.
UID ID for this user
Username (*) Username for this user
Admin Org ID ID of the administrator organization
Admin Org Name Name of the administrator's organization
Org ID ID of the user's organization
Org Name (*) Name of the user's organization
Group ID ID for this user's group
Group Name Name of the user's group
Admin UID ID for this administrator (if applicable)
Admin Username Username for this administrator (if applicable)
Old values (*) User account settings before the administrator made changes. For example, if the administrator activity changed the workers for the related user, this field shows the "max_workers" value before the change.
New values (*) User account settings after the administrator made changes. For example, if the administrator activity changed the workers for the related user, this field shows the "max_workers" value after the change.

(*) denotes an identifying field for this report

App Usage activity report

The App Usage activity output reports the following data about application events.

Report field Description
Timestamp—UTC Timestamp (UTC time standard) when the application event occurred
Event Type of application event, such as Add Model, Compliance Doc Generated, aiAPI Portal Login, Dataset Upload, etc.
UID ID of the user
Username (*) Name of the user
Project ID ID of the project
Project Name (*) Name of the project
Org ID ID of the user's organization
Org Name (*) Name of the user's organization
Group ID ID of the user's group
Group Name (*) Name of the user's group
User Role Role for the user who initiated the event; values include OWNER, USER, OBSERVER
Project Type Type for the related project; possible values include Binary Classification, Regression, Time Series—Regression, Multiclass Classification, etc.
Metric Optimization metric for the related project; potential values include LogLoss, RMSE, AUC, etc.
Partition Method Partition method for the related project (i.e., how data is partitioned for this project)
Target Variable (*) Target variable for the related project (i.e., what DataRobot will predict)
Model ID ID of the model
Model Type Type for the model; this also is the name of the model or blueprint
Blender Model Types Type of blender model, if applicable
Sample Type Type of training sample for the project; values may include Sample Percent, Row Count, Duration, etc.
Sample Length Amount of sample data for training the project; values are based on Sample Type and may be percentage, number of rows, or length of time
Model Fit Time Amount of time (in seconds) used to build the model
Recommended Model Identifies if this is the recommended model for deployment (true) or not (false)
Insight Type Type of insight requested for this model; possible values may include Variable Importance, Compute Series Accuracy, Compute Accuracy Over Time, Dual Lift, etc.
Custom Template Identifies if the compliance document (for the event) was developed with a custom template (true) or not (false); applies to Compliance Doc Generated events
Deployment ID ID for the deployment; applies to Replaced Model events
Deployment Type Type of deployment; applies to events such as Replaced Model and Deployment Added, and possible values include Dedicated Prediction (deployment to a dedicated prediction server) or Secure Worker (in-app modeling workers used for predictions)
Client Type Client (DataRobotPythonClient or DataRobotRClient) used to interface with DataRobot; applies to events such as DataSet Upload, Project Created, Project Target Selected, Select Model Metric, etc.
Client Version Version of the related client (DataRobotPythonClient or DataRobotRClient)
Catalog ID The ID of an item in the catalog. It can be used to address an individual item. Catalog ID is the same as a Dataset ID when the catalog item is a dataset.
Catalog Version ID An ID indicating a specific version of a catalog item. Catalog items can have multiple versions; by default the catalog uses the latest version. To work with an earlier version, you must use the specific version ID.
Dataset ID ID of dataset
Dataset Name (*) Name of dataset
Dataset Size Size of project dataset
Snapshotted State A dataset can be snapshotted (materialized) or not. If it is snapshotted, the data has been stored locally. If it is not, the data is requested from the source whenever it is used.
Grantee The UID of the user sharing the asset
With Grant Indicates whether the user receiving the new role is allowed to share with others

(*) denotes an identifying field for this report

Prediction Usage activity report

The Prediction Usage activity output reports the following data about prediction statistics.

Report field Description
Timestamp—UTC Timestamp (UTC time standard) when the prediction event occurred
UID ID for the user who initiated this event
Username (*) Name for the user who initiated this event
Project ID ID of the project
Org ID ID of the user's organization
Org Name (*) Name of the user's organization
Group ID ID of the user's group
Group Name (*) Name of the user's group
User Role Role for the user who initiated the prediction event; values include OWNER, USER, OBSERVER
Model ID ID of the model
Model Type Type for the model; this also is the name of the model or blueprint
Blender Model Types Type of blender model, if applicable
Recommended Model Identifies if this is the recommended model for deployment (true) or not (false)
Project Type Type for the project; possible values include Binary Classification, Regression, Time Series—Regression, Multiclass Classification, etc.
Deployment ID ID of the deployment
Deployment Type Type of deployment; possible values include dedicated (deployment to a dedicated prediction server) or Secure Worker (in-app modeling workers used for predictions)
Dataset ID ID of the dataset
Prediction Method Method for making predictions for the related project; possible values are Modeling Worker (predictions using modeling workers) or Dedicated Prediction (predictions using dedicated prediction server)
Prediction Explanations Identifies if Prediction Explanations were computed for this project model (true) or not (false)
# of Requests Number of prediction requests the deployment has received (where a single request can contain multiple prediction requests); provides statistics for deployment service health
# Rows Scored Number of dataset rows scored (for predictions) for this deployment
User Errors Number of user errors (4xx errors) for this deployment; provides statistics for deployment service health
Server Errors Number of server errors (5xx errors) for this deployment; provides statistics for deployment service health
Average Execution Time Average time (in milliseconds) DataRobot spent processing prediction requests for this deployment; provides statistics for deployment service health

(*) denotes an identifying field for this report


Updated November 30, 2021
Back to top