Dataiku Agent Management - Reference documentation

Welcome to the reference documentation for Dataiku Agent Management. This documentation contains:

• The main documentation of the concepts, interfaces and features of Dataiku Agent Management

• Information on how to install and configure Dataiku Agent Management

• Information for administrators on how to operate Dataiku Agent Management

You might also find these other resources useful:

• The Dataiku Community is a place where you can join the discussion, get support, share best practices and engage with other Dataiku users.

• Concepts
  • Agent Management concepts
  • Getting started with Agent Management
• Platforms
  • Creating a platform
  • Scanning and discovery
  • Platform activity
  • Platform logs
  • Platform settings
  • Uptime tests
• Supported platforms
  • Capability summary
  • Platform reference
• Inventory
  • Finding agents
  • Enriching an agent
• Monitoring
  • Agent overview
  • Operational and usage metrics
  • KPI metrics
  • Cost
  • Tools and agent properties
  • Agent Reviews
  • Dashboard controls
  • Activity and collection health
• KPIs
  • Designing a KPI Flow
  • Required dataset schema
  • Configuring an agent KPI
  • Inline and global KPIs
  • Interpreting KPI values
• Cost Management
  • Model pricing
  • Portfolio cost
  • Agent cost
• Risk Management
  • Associating risks
  • Assessing a risk
  • Sign-off
  • Portfolio risk
• Alerting
  • Alerting targets
  • Metric trigger conditions
  • Agent Review trigger conditions
  • Recovery, acknowledgement, and closure
  • Reporters and notifications
  • Reviewing and managing alerts
• Dataiku AI in Agent Management
  • Overview and setup
  • Cobuild
  • Usage of AI Services
• Data Management concepts
  • Data
  • Datasets
  • Recipes
  • Building datasets
  • Managed and external datasets
  • Partitioning
• Connecting to data
  • Supported connections
  • File formats
  • Snowflake
  • Databricks
  • Google BigQuery
  • Amazon Redshift
  • Azure Synapse
  • Microsoft Fabric Warehouse
  • Other SQL databases
  • Amazon S3
  • Azure Blob Storage
  • Google Cloud Storage
  • Iceberg
  • Upload your files
  • Elasticsearch and OpenSearch
  • Application connectors
  • Managed folders
  • “Files in folder” dataset
  • Metrics dataset
  • Internal stats dataset
  • “Editable” dataset
  • kdb+
  • Server filesystem
  • Dataset plugins
  • OData
  • Making relocatable managed datasets
  • Data ordering
• Exploring data
  • Sampling
  • Analyze
• The Flow
  • Visual Grammar
  • Flow zones
  • Rebuilding Datasets
  • Inserting and deleting recipes
  • Limiting Concurrent Executions
  • Exporting the Flow to PDF or images
  • How to Manage Large Flows with Flow Folding
  • Flow Document Generator
  • Flow explanations
• Data preparation
  • How to Copy Prepare Recipe Steps
  • Sampling
  • Execution engines
  • Processors reference
  • Filtering and flagging rows
  • Managing dates
  • Reshaping
  • Geographic processors
• Visual recipes
  • Prepare: Cleanse, Normalize, and Enrich
  • Sync: copying datasets
  • Grouping: aggregating data
  • Window: analytics functions
  • Distinct: get unique rows
  • Join: joining datasets
  • Fuzzy join: joining two datasets
  • Geo join: joining datasets based on geospatial features
  • Splitting datasets
  • Top N: retrieve first N rows
  • Stacking datasets
  • Sampling datasets
  • Sort: order values
  • Pivot recipe
  • Extracting document content
  • Generate statistics
  • Push to editable recipe
  • Download recipe
  • List Folder Contents
  • Dynamic recipe repeat
  • AI Assistant: Generate Recipe feature
  • Extract failed rows
  • Upsert: Consolidate data
  • List Access
• Code recipes
  • The common editor layout
  • Python recipes
  • SQL recipes
  • Variables expansion in code recipes
  • Dynamic recipe repeat
  • Code explanations
• Generative AI and LLM Mesh
  • Introduction
  • LLM connections
  • Running Hugging Face models
  • Prompt Studios
  • Prompt Recipe
  • Adding Knowledge to LLMs
  • Guardrails
  • LLM Mesh API
  • Multimodal capabilities
  • Fine-tuning
  • Cost Control
  • Rate Limiting
  • Evaluating LLMs & GenAI use cases
  • Automated Prompt Optimization
• AI Agents
  • Introduction to Agents in Dataiku
  • External Agents
  • Agent Review
  • Tracing
  • Agent Evaluation
• Semantic Models
  • Create, import and manage Semantic Models
  • Semantic Model Query Tool
  • Dataset SQL Query Tool
• Charts
  • The Charts Interface
  • Sampling & Engine
  • Basic Charts
  • Tables
  • Scatter Charts
  • Map Charts
  • Other Charts
  • Common Chart Elements
  • Color palettes
  • Formatting
  • Sorting
  • Filter settings
  • Custom aggregations
  • Reusable dimensions
  • Hierarchies and drill down
  • Reference lines
  • Copy/paste charts
  • Third-party BI tools
  • Tableau Hyper Export
  • Microsoft Power BI v2
  • QLIK QVX
  • MicroStrategy
• Schemas, storage types and meanings
  • Definitions
  • Basic usage
  • Schema for data preparation
  • Creating schemas of datasets
  • Handling of schemas by recipes
  • List of recognized meanings
  • User-defined meanings
  • Handling and display of dates
• Code notebooks
  • SQL notebook
  • Python notebooks
  • Installing Jupyter Extensions
• Collaboration
  • Wikis
  • Discussions
  • Markdown
  • Tags
  • Working with Git
  • Version control of projects
• Catalog
  • Catalog Collections
  • Most Used Datasets
  • Assets
• Data Lineage
  • How to access
  • Manual lineage
  • Exporting lineage graph to PDF or images
• Working with partitions
  • Partitioning files-based datasets
  • Partitioned SQL datasets
  • Specifying partition dependencies
  • Partition identifiers
  • Recipes for partitioned datasets
  • Partitioned Hive recipes
  • Partitioned SQL recipes
  • Partitioning variables substitutions
  • Partitioned Models
  • The two partitioning models
• Agent Management and SQL
  • SQL datasets
  • SQL write and execution
  • Partitioning
  • SQL pipelines in Agent Management
• Agent Management and Python
  • Installing Python packages
  • Reusing Python code
  • Using Matplotlib
  • Using SpaCy
  • Using Bokeh
  • Using Plot.ly
  • Using Ggplot
  • Using Jupyter Widgets
  • Documentation Portal
• Code environments
  • Operations (Python)
  • Base packages
  • Plugins’ code environments
  • Custom options and environment
  • Troubleshooting
  • Code env permissions
• Metrics, checks and Data Quality
  • Metrics
  • Checks
  • Data Quality Rules
  • Custom probes and checks
  • Data Quality Templates
• Automation scenarios
  • Definitions
  • Scenario steps
  • Launching a scenario
  • Reporting on scenario runs
  • Custom scenarios
  • Variables in scenarios
  • Step-based execution control
  • Testing a project
• Python APIs
• Public REST API
  • Features
  • Public API Keys
  • The REST API
• Installing and setting up
  • Dataiku Cloud Stacks for AWS
  • Dataiku Cloud Stacks for Azure
  • Dataiku Cloud Stacks for GCP
  • Custom Dataiku install on Linux
• Operating Agent Management
  • Agent Management global settings
  • dsscli tool
  • The data directory
  • Backing up
  • Audit trail
  • The runtime databases
  • Logging in Dataiku Agent Management
  • Dataiku Agent Management Macros
  • Managing Dataiku Agent Management disk usage
  • Understanding and tracking Agent Management processes
  • Tuning and controlling memory usage
  • Using cgroups for resource control
  • Monitoring Agent Management
  • HTTP proxies
  • Agent Management license
  • Compute resource usage reporting
• Security
  • Agent Management permissions
  • Authentication
  • Project Access
  • Project folders - defaults and access
  • Permissions
  • Connections security
  • User profiles
  • Shared objects
  • Authorized objects
  • Secrets Manager
  • User secrets
  • Audit Trail
  • Passwords security
  • Advanced security options
  • Messaging Channels Permissions
  • How Dataiku responds to the October 2026 vulnerabilities
• User Isolation
  • Capabilities of User Isolation Framework
  • Prerequisites and limitations
  • Initial Setup
• Formula language
  • Basic usage
  • Reading column values
  • Variables typing and autotyping
  • Boolean values
  • Operators
  • Array and object operations
  • Object notations
  • Dataiku Agent Management variables
  • Array functions
  • Boolean functions
  • Date functions
  • Math functions
  • Object functions
  • String functions
  • Geometry functions
  • Value access functions
  • Control structures
  • Tests
• Custom variables expansion
  • Defining variables
  • Using variables in the code of a recipe
  • Using variables in configuration fields
  • Using override tables
  • Using variables in a dashboard
  • Modifying the value of variables
  • Using filters
• Sampling methods
  • Generic sampling methods
  • Sampling methods availability
• Accessibility
  • Global Shortcuts
  • Project Navigation
  • Within the Flow
  • Within a Dataset
  • Within a Prepare Recipe
  • Within a Code Recipe
  • Within any Recipe
  • Within any Code Editor (Excluding Notebooks)
  • Within any Flow Object
  • Within Plugins Development
  • Within a Dataset Insight
  • Within a Dashboard
• Troubleshooting
  • Diagnosing and debugging issues
  • Obtaining support
  • Support tiers
  • Common issues
  • Error codes
• Release notes
  • Agent Management 15 release notes
• Other Documentation
  • Other Dataiku products
• Third-party acknowledgements
  • Agent Management
  • Fleet Manager
  • Mac version only
  • Dataiku Cloud
  • AI Services Third Party AI Providers
  • Stories
  • Dataiku Solutions
  • Dataiku Plugins