Configuration for anomaly detection on a ReporterResource. Admin defines what to detect; system generates the execution plan.

An asterisk * in the title indicates the field is mandatory.
If a field has a default value, it is shown in the Description column.
If the field type is an array, the Field Name has a .[n] suffix, where n is the array index placeholder.
Object and array field names are listed to provide context. If a field belongs to an object or an array, the full field name is shown in dot separated notation.

Title						Description	Details
Configuration						Group Assigned by FDP	Field Name: Configuration Type: Object
	Name *					Unique name for this anomaly detection configuration	Field Name: Configuration.name Type: String MaxLength: 1024
	Detection Question *					What the admin wants to detect	Field Name: Configuration.detection_question Type: String MaxLength: 1024
	Resource *					ReporterResource to monitor for anomalies	Field Name: Configuration.resource Type: String Target: data/ReporterResource Target attr: name MaxLength: 1024 Format: uri
	Data Source Override					Override the default data source for this configuration (business key reference)	Field Name: Configuration.datasource_override Type: String Target: data/DashboardDatasource MaxLength: 1024 Format: uri
	Run Summary Retention (Days)					Number of days to retain detection run summary records before automatic purge. Applies to new runs only — changing this value does not affect existing run records. Default: 30	Field Name: Configuration.run_summary_retention_days Type: Integer Default: 30 Choices: ["7 Days", "14 Days", "30 Days", "60 Days", "90 Days", "180 Days", "365 Days"]
		Features				Optional Postgres column names to analyze. When empty, the LLM derives optimal features from the resource schema and detection question.	Field Name: features.[n] Type: Array
Detection						Group Assigned by FDP	Field Name: Detection Type: Object
		Desired State Search				Ephemeral search filters to find a source instance from the resource's base table. Results populate desired_state. Not persisted on save.	Field Name: desired_state_search.[n] Type: Array
				Filter Options			Field Name: filter_options.[n] Type: Array
					Filter By		Field Name: Detection.desired_state_search.[n].filter_options.[n].filter_by Type: String MaxLength: 1024
					Filter Type		Field Name: Detection.desired_state_search.[n].filter_options.[n].filter_type Type: String MaxLength: 1024 Choices: ["Contains", "Does Not Contain", "Starts With", "Ends With", "Equals", "Not Equal"]
					Filter String		Field Name: Detection.desired_state_search.[n].filter_options.[n].filter_string Type: String MaxLength: 1024
					Ignore Case		Field Name: Detection.desired_state_search.[n].filter_options.[n].ignore_case Type: Boolean
	Desired State Instance					Ephemeral dropdown showing instances matching the search filters. Selecting an instance populates desired_state. Not persisted on save.	Field Name: Detection.desired_state_instance Type: String MaxLength: 1024
		Desired State				Locked-in key-value pairs representing the desired configuration state for drift detection	Field Name: desired_state.[n] Type: Array
			Field *			Resource property name — column in the Postgres base table	Field Name: Detection.desired_state.[n].field Type: String MaxLength: 1024
			Value *			Desired value for this property	Field Name: Detection.desired_state.[n].value Type: String MaxLength: 1024
		Detection Filter				Array of filter groups (OR'd) to scope detection to a subset of rows. Each group contains a filter_options array of filter items (AND'd within the group).	Field Name: detection_filter.[n] Type: Array
			Name			Optional label for this filter group.	Field Name: Detection.detection_filter.[n].name Type: String MaxLength: 1024
			Description			Optional description explaining the purpose of this filter group.	Field Name: Detection.detection_filter.[n].description Type: String MaxLength: 1024
				Filter Options			Field Name: filter_options.[n] Type: Array
					Filter By		Field Name: Detection.detection_filter.[n].filter_options.[n].filter_by Type: String MaxLength: 1024
					Filter Type		Field Name: Detection.detection_filter.[n].filter_options.[n].filter_type Type: String MaxLength: 1024 Choices: ["Contains", "Does Not Contain", "Starts With", "Ends With", "Equals", "Not Equal"]
					Filter String		Field Name: Detection.detection_filter.[n].filter_options.[n].filter_string Type: String MaxLength: 1024
					Ignore Case		Field Name: Detection.detection_filter.[n].filter_options.[n].ignore_case Type: Boolean
Execution Plan						Group Assigned by FDP	Field Name: Execution Plan Type: Object
	Allow Playbook Edit					Enable expert mode to manually edit the execution plan. When enabled, operational fields (tool, parameters, exit on failure) become editable. Changes may be overwritten when the plan is regenerated.	Field Name: Execution Plan.allow_playbook_edit Type: Boolean
	Playbook Configuration					Structured execution plan generated by the generate_execution_plan operation. Contains required datasources, ordered detection steps, and LLM reasoning. Cleared automatically when resource or features change.	Field Name: playbook_config Type: Object
			Required Datasources			List of datasource types needed by this plan (e.g. 'automate-postgres'). Empty array when datasource_override is set.	Field Name: required_datasources.[n] Type: Array
		Reasoning				LLM-generated explanation of why this detection approach was chosen.	Field Name: Execution Plan.playbook_config.reasoning Type: String MaxLength: 1024
			Steps			Ordered list of detection steps to execute. Each step invokes a tool with specific parameters.	Field Name: steps.[n] Type: Array
				Tool *		Name of the detection tool to execute in this step.	Field Name: Execution Plan.playbook_config.steps.[n].tool Type: String MaxLength: 1024 Choices: ["Query Snapshot Data", "Detect Config Drift", "ML Detect Outliers", "Query Reporter Metrics", "ML Time Series Analysis", "Create Anomaly", "Summarize Detection Results"]
				Parameters		Tool-specific parameters as a JSON object. Structure varies per tool.	Field Name: Execution Plan.playbook_config.steps.[n].params Type: String MaxLength: 1024
				Goal *		Human-readable description of what this step accomplishes.	Field Name: Execution Plan.playbook_config.steps.[n].goal Type: String MaxLength: 1024
				Rationale		Explanation of why this step is needed in the detection plan.	Field Name: Execution Plan.playbook_config.steps.[n].rationale Type: String MaxLength: 1024
				Exit on Failure		If true, abort the entire detection run when this step fails. If false, log the error and continue to the next step.	Field Name: Execution Plan.playbook_config.steps.[n].exit_on_failure Type: Boolean
		Presentation Schema				Defines the sections and structure for the LLM-generated run summary. The summary is written for non-ML-expert admins. Customize sections to control what appears in the report.	Field Name: presentation_schema Type: Object
			Type *			The type of presentation format	Field Name: Execution Plan.playbook_config.presentation_schema.type Type: String MaxLength: 1024 Choices: ["Narrative - Structured narrative format with sections", "Structured JSON - Machine-readable output format"]
				Sections		List of sections in the presentation	Field Name: sections.[n] Type: Array
					Header *	Section header/title	Field Name: Execution Plan.playbook_config.presentation_schema.sections.[n].header Type: String MaxLength: 1024
					Description	Description of what this section contains	Field Name: Execution Plan.playbook_config.presentation_schema.sections.[n].description Type: String MaxLength: 1024
			Note			Additional notes about the presentation format	Field Name: Execution Plan.playbook_config.presentation_schema.note Type: String MaxLength: 1024

Model: data/AnomalyDetectionConfig

Model Details: data/AnomalyDetectionConfig