openapi: 3.0.3
info:
  title: Keeper Data Bridge API
  description: |
    REST API for the Keeper Data Bridge backend service.
    Provides endpoints for data import, OData-style querying, cleanse analysis,
    issue management, ad-hoc report export, holdings lookup, throttle policy management,
    and MongoDB performance benchmarking with noisy-neighbour diagnostics.
  version: 1.0.0
  contact:
    name: DEFRA

servers:
  - url: /api
    description: Base API path

tags:
  - name: Import
    description: Bulk data import, reporting, lineage, collection management, and storage operations.
  - name: Query
    description: OData-style querying of MongoDB collections.
  - name: Cleanse
    description: Cleanse analysis operations, issue management, and report export.
  - name: Cleanse Export
    description: |
      Ad-hoc full cleanse report export operations.
      Exports all active issues to CSV/ZIP on S3 with progress tracking.
      Runs independently of the analysis pipeline — does not send notifications
      or update the incremental-export timestamp.
  - name: Holdings
    description: CTS and SAM CPH holding lookups.
  - name: Throttle Policies
    description: Throttle policy CRUD and activation.
  - name: External Catalogue
    description: External file catalogue browsing and file uploads.
  - name: Benchmark
    description: |
      Self-contained MongoDB benchmark for diagnosing environment-level
      performance differences. Creates temporary collections, runs read/write
      scenarios, and produces a detailed report with noisy-neighbour analysis.
      Does not touch production data collections.

paths:
  # ═══════════════════════════════════════════════════════════════════
  # Import
  # ═══════════════════════════════════════════════════════════════════
  /import/start:
    post:
      tags: [Import]
      operationId: startBulkImport
      summary: Start a bulk file import
      description: |
        Starts a bulk file import process asynchronously.
        Returns immediately with an import ID once the lock has been acquired.
        The import continues running in the background.
      parameters:
        - name: sourceType
          in: query
          description: "The source type for the import ('internal' or 'external')"
          schema:
            type: string
            default: external
            enum: [internal, external]
      responses:
        '202':
          description: Import started successfully
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/StartBulkImportResponse'
        '400':
          description: Invalid sourceType parameter
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/ErrorResponse'
        '409':
          description: Another import is already running
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/ErrorResponse'
        '499':
          description: Client closed request
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/ErrorResponse'

  /import:
    get:
      tags: [Import]
      operationId: getImportSummaries
      summary: Get paginated import summaries
      description: Returns import summaries in reverse chronological order.
      parameters:
        - $ref: '#/components/parameters/Skip'
        - $ref: '#/components/parameters/Top'
      responses:
        '200':
          description: Paginated list of import summaries
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/ImportSummariesResponse'
        '400':
          description: Invalid pagination parameters
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/ErrorResponse'

  /import/{importId}:
    get:
      tags: [Import]
      operationId: getImportReport
      summary: Get an import report
      description: Gets the full import report for a specific import ID.
      parameters:
        - name: importId
          in: path
          required: true
          schema:
            type: string
            format: uuid
      responses:
        '200':
          description: Import report
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/ImportReport'
        '404':
          description: Import not found

  /import/{importId}/files:
    get:
      tags: [Import]
      operationId: getFileReports
      summary: Get file reports for an import
      description: Returns file processing reports for all files in a specific import.
      parameters:
        - name: importId
          in: path
          required: true
          schema:
            type: string
            format: uuid
      responses:
        '200':
          description: File reports
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/FileReportsResponse'
        '404':
          description: Import not found

  /import/lineage/{collectionName}/{recordId}:
    get:
      tags: [Import]
      operationId: getRecordLineageEvents
      summary: Get record lineage events
      description: |
        Gets paginated lineage events for a specific record in chronological order.
        The recordId is a URL-safe SHA256 hash generated from composite key parts.
      parameters:
        - name: collectionName
          in: path
          required: true
          description: "The collection name (e.g., 'sam_cph_holdings')"
          schema:
            type: string
        - name: recordId
          in: path
          required: true
          description: URL-safe record ID (SHA256 hash)
          schema:
            type: string
        - $ref: '#/components/parameters/Skip'
        - $ref: '#/components/parameters/Top'
      responses:
        '200':
          description: Paginated lineage events
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/PaginatedLineageEvents'
        '400':
          description: Invalid parameters
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/ErrorResponse'
        '404':
          description: Record lineage not found
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/ErrorResponse'
        '499':
          description: Client closed request
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/ErrorResponse'

  /import/generate-record-id:
    post:
      tags: [Import]
      operationId: generateRecordId
      summary: Generate a record ID from key parts
      description: Generates a URL-safe record ID from composite key parts using SHA256 hashing.
      requestBody:
        required: true
        content:
          application/json:
            schema:
              $ref: '#/components/schemas/GenerateRecordIdRequest'
      responses:
        '200':
          description: Generated record ID
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/GenerateRecordIdResponse'
        '400':
          description: Invalid key parts
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/ErrorResponse'

  /import/collections/{collectionName}:
    delete:
      tags: [Import]
      operationId: deleteCollection
      summary: Delete a MongoDB collection
      description: Deletes a specific MongoDB collection by name. Must be defined in DataSetDefinitions.
      parameters:
        - name: collectionName
          in: path
          required: true
          schema:
            type: string
      responses:
        '200':
          description: Collection deleted
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/DeleteCollectionResponse'
        '404':
          description: Collection not found in definitions
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/ErrorResponse'
        '499':
          description: Client closed request
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/ErrorResponse'

  /import/collections:
    delete:
      tags: [Import]
      operationId: deleteAllCollections
      summary: Delete all MongoDB collections
      description: Deletes all MongoDB collections defined in DataSetDefinitions.
      responses:
        '200':
          description: Collections deleted
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/DeleteCollectionsResponse'
        '499':
          description: Client closed request
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/ErrorResponse'

  /import/reporting-collections/{collectionName}:
    delete:
      tags: [Import]
      operationId: deleteReportingCollection
      summary: Delete a reporting collection
      description: |
        Deletes a specific reporting/lineage collection.
        Valid collections: import_reports, import_files, record_lineage, record_lineage_events.
      parameters:
        - name: collectionName
          in: path
          required: true
          schema:
            type: string
      responses:
        '200':
          description: Reporting collection deleted
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/DeleteCollectionResponse'
        '404':
          description: Collection not valid
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/ErrorResponse'
        '499':
          description: Client closed request
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/ErrorResponse'

  /import/reporting-collections:
    delete:
      tags: [Import]
      operationId: deleteAllReportingCollections
      summary: Delete all reporting collections
      description: |
        Deletes all reporting/lineage collections.
        Includes: import_reports, import_files, record_lineage, record_lineage_events.
      responses:
        '200':
          description: Reporting collections deleted
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/DeleteCollectionsResponse'
        '499':
          description: Client closed request
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/ErrorResponse'

  /import/internal-storage:
    delete:
      tags: [Import]
      operationId: clearDownInternalStorage
      summary: Clear down internal storage
      description: Deletes all objects from internal target storage under the configured prefix.
      parameters:
        - name: sourceType
          in: query
          schema:
            type: string
            default: internal
            enum: [internal, external]
      responses:
        '200':
          description: Storage cleared
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/ClearDownStorageResponse'
        '499':
          description: Client closed request
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/ErrorResponse'

  # ═══════════════════════════════════════════════════════════════════
  # Query
  # ═══════════════════════════════════════════════════════════════════
  /query/{collectionName}:
    get:
      tags: [Query]
      operationId: queryCollection
      summary: Query a MongoDB collection
      description: |
        Query a MongoDB collection with OData-style parameters.

        Examples:
        - `GET /api/query/sam_cph_holdings?$filter=CPH eq 'ABC123'&$top=50`
        - `GET /api/query/sam_cph_holdings?$filter=contains(CPH,'ABC') and IsDeleted eq false&$orderby=UpdatedAtUtc desc`
        - `GET /api/query/sam_cph_holdings?$select=CPH,UpdatedAtUtc&$top=20`
      parameters:
        - name: collectionName
          in: path
          required: true
          description: "Name of the collection (e.g., 'sam_cph_holdings')"
          schema:
            type: string
        - name: $filter
          in: query
          description: OData $filter expression
          schema:
            type: string
        - name: $orderby
          in: query
          description: "OData $orderby expression (e.g., 'UpdatedAtUtc desc')"
          schema:
            type: string
        - name: $select
          in: query
          description: "OData $select expression (e.g., 'CPH,UpdatedAtUtc')"
          schema:
            type: string
        - name: $skip
          in: query
          schema:
            type: integer
        - name: $top
          in: query
          description: Max 1000, default 100
          schema:
            type: integer
            default: 100
            maximum: 1000
        - name: $count
          in: query
          description: Include total count in response
          schema:
            type: boolean
            default: true
      responses:
        '200':
          description: Query results
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/QueryResult'
        '400':
          description: Invalid query parameters
        '404':
          description: Collection not found
        '499':
          description: Client closed request
        '500':
          description: Internal server error

  # ═══════════════════════════════════════════════════════════════════
  # Cleanse — Operations
  # ═══════════════════════════════════════════════════════════════════
  /cleanse/start-analysis:
    post:
      tags: [Cleanse]
      operationId: startAnalysis
      summary: Start a cleanse analysis
      description: |
        Starts a new cleanse analysis operation in the background.
        Progress can be tracked via GET /cleanse/run/{operationId}.
      responses:
        '202':
          description: Analysis started
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/StartAnalysisResponse'
        '409':
          description: An analysis is already running
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/ErrorResponse'
        '499':
          description: Client closed request
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/ErrorResponse'

  /cleanse/cancel-analysis:
    post:
      tags: [Cleanse]
      operationId: cancelAnalysis
      summary: Cancel the running analysis
      description: Requests cancellation of the currently running cleanse analysis operation.
      responses:
        '200':
          description: Cancellation requested
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/CancelAnalysisResponse'
        '404':
          description: No running analysis found
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/ErrorResponse'
        '499':
          description: Client closed request
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/ErrorResponse'

  /cleanse/runs:
    get:
      tags: [Cleanse]
      operationId: getAnalysisRuns
      summary: Get paginated analysis run summaries
      parameters:
        - $ref: '#/components/parameters/Skip'
        - name: top
          in: query
          schema:
            type: integer
            default: 10
            minimum: 1
            maximum: 100
      responses:
        '200':
          description: Analysis run summaries
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/AnalysisRunsResponse'
        '400':
          description: Invalid pagination parameters
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/ErrorResponse'
        '499':
          description: Client closed request
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/ErrorResponse'

  /cleanse/run/{operationId}:
    get:
      tags: [Cleanse]
      operationId: getAnalysisRun
      summary: Get full details of an analysis run
      description: |
        Returns the complete operation state including per-phase progress,
        live RPM statistics, and projected completion times.
      parameters:
        - name: operationId
          in: path
          required: true
          schema:
            type: string
      responses:
        '200':
          description: Analysis run details
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/CleanseAnalysisOperationDto'
        '404':
          description: Operation not found
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/ErrorResponse'
        '499':
          description: Client closed request
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/ErrorResponse'

  /cleanse/run/{operationId}/regenerate-url:
    post:
      tags: [Cleanse]
      operationId: regenerateReportUrl
      summary: Regenerate a report presigned URL
      description: Regenerates the presigned S3 URL for an operation's report when the original has expired.
      parameters:
        - name: operationId
          in: path
          required: true
          schema:
            type: string
      responses:
        '200':
          description: URL regenerated
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/RegenerateUrlResponse'
        '400':
          description: Operation has no report
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/ErrorResponse'
        '404':
          description: Operation not found
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/ErrorResponse'
        '499':
          description: Client closed request
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/ErrorResponse'

  /cleanse/delete-data:
    post:
      tags: [Cleanse]
      operationId: deleteCleanseData
      summary: Delete all cleanse issues
      responses:
        '200':
          description: Data deleted
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/DeleteDataResponse'
        '499':
          description: Client closed request
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/ErrorResponse'
        '500':
          description: Server error
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/ErrorResponse'

  /cleanse/delete-metadata:
    post:
      tags: [Cleanse]
      operationId: deleteCleanseMetadata
      summary: Delete all operation history
      responses:
        '200':
          description: Metadata deleted
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/DeleteDataResponse'
        '499':
          description: Client closed request
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/ErrorResponse'
        '500':
          description: Server error
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/ErrorResponse'

  /cleanse/test-notification:
    post:
      tags: [Cleanse]
      operationId: sendTestNotification
      summary: Send a test notification email
      description: Sends a test email via GOV.UK Notify to verify configuration.
      responses:
        '200':
          description: Test notification sent
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/TestNotificationResponse'
        '499':
          description: Client closed request
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/ErrorResponse'
        '500':
          description: Notification failed
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/ErrorResponse'

  # ═══════════════════════════════════════════════════════════════════
  # Cleanse — Issues
  # ═══════════════════════════════════════════════════════════════════
  /cleanse/issues:
    get:
      tags: [Cleanse]
      operationId: getIssues
      summary: Get paginated, filterable cleanse issues
      parameters:
        - $ref: '#/components/parameters/Skip'
        - $ref: '#/components/parameters/Top'
        - name: ctsLidFullIdentifier
          in: query
          description: Filter by CTS LID (contains, case-insensitive)
          schema:
            type: string
        - name: cph
          in: query
          description: Filter by CPH (contains, case-insensitive)
          schema:
            type: string
        - name: issueCode
          in: query
          description: Filter by exact issue code
          schema:
            type: string
        - name: ruleCode
          in: query
          description: Filter by exact rule code
          schema:
            type: string
        - name: errorCode
          in: query
          description: Filter by exact error code
          schema:
            type: string
        - name: isActive
          in: query
          schema:
            type: boolean
        - name: isIgnored
          in: query
          schema:
            type: boolean
        - name: resolutionStatus
          in: query
          description: "None, Todo, InProgress, Resolved"
          schema:
            type: string
            enum: [None, Todo, InProgress, Resolved]
        - name: assignedTo
          in: query
          schema:
            type: string
        - name: isUnassigned
          in: query
          schema:
            type: boolean
        - name: createdAfterUtc
          in: query
          schema:
            type: string
            format: date-time
        - name: createdBeforeUtc
          in: query
          schema:
            type: string
            format: date-time
        - name: updatedAfterUtc
          in: query
          schema:
            type: string
            format: date-time
        - name: updatedBeforeUtc
          in: query
          schema:
            type: string
            format: date-time
        - name: sortBy
          in: query
          description: Field to sort by (default LastUpdatedAtUtc)
          schema:
            type: string
        - name: sortDescending
          in: query
          schema:
            type: boolean
            default: true
      responses:
        '200':
          description: Paginated issues
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/IssuesResponse'
        '400':
          description: Invalid parameters
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/ErrorResponse'
        '499':
          description: Client closed request
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/ErrorResponse'

  /cleanse/issues/{issueId}/ignore:
    post:
      tags: [Cleanse]
      operationId: ignoreIssue
      summary: Ignore an issue
      parameters:
        - name: issueId
          in: path
          required: true
          schema:
            type: string
      requestBody:
        required: true
        content:
          application/json:
            schema:
              $ref: '#/components/schemas/PerformedByRequest'
      responses:
        '200':
          description: Issue ignored
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/IssueCommandResponse'
        '404':
          description: Issue not found
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/ErrorResponse'
        '499':
          description: Client closed request
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/ErrorResponse'

  /cleanse/issues/{issueId}/unignore:
    post:
      tags: [Cleanse]
      operationId: unignoreIssue
      summary: Remove ignored flag from an issue
      parameters:
        - name: issueId
          in: path
          required: true
          schema:
            type: string
      requestBody:
        required: true
        content:
          application/json:
            schema:
              $ref: '#/components/schemas/PerformedByRequest'
      responses:
        '200':
          description: Issue unignored
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/IssueCommandResponse'
        '404':
          description: Issue not found
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/ErrorResponse'
        '499':
          description: Client closed request
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/ErrorResponse'

  /cleanse/issues/{issueId}/resolution-status:
    post:
      tags: [Cleanse]
      operationId: updateResolutionStatus
      summary: Update resolution status
      parameters:
        - name: issueId
          in: path
          required: true
          schema:
            type: string
      requestBody:
        required: true
        content:
          application/json:
            schema:
              $ref: '#/components/schemas/UpdateResolutionStatusRequest'
      responses:
        '200':
          description: Status updated
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/IssueCommandResponse'
        '400':
          description: Invalid status value
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/ErrorResponse'
        '404':
          description: Issue not found
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/ErrorResponse'
        '499':
          description: Client closed request
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/ErrorResponse'

  /cleanse/issues/{issueId}/assign:
    post:
      tags: [Cleanse]
      operationId: assignIssue
      summary: Assign an issue to a user
      parameters:
        - name: issueId
          in: path
          required: true
          schema:
            type: string
      requestBody:
        required: true
        content:
          application/json:
            schema:
              $ref: '#/components/schemas/AssignIssueRequest'
      responses:
        '200':
          description: Issue assigned
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/IssueCommandResponse'
        '404':
          description: Issue not found
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/ErrorResponse'
        '499':
          description: Client closed request
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/ErrorResponse'

  /cleanse/issues/{issueId}/unassign:
    post:
      tags: [Cleanse]
      operationId: unassignIssue
      summary: Unassign an issue
      parameters:
        - name: issueId
          in: path
          required: true
          schema:
            type: string
      requestBody:
        required: true
        content:
          application/json:
            schema:
              $ref: '#/components/schemas/PerformedByRequest'
      responses:
        '200':
          description: Issue unassigned
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/IssueCommandResponse'
        '404':
          description: Issue not found
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/ErrorResponse'
        '499':
          description: Client closed request
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/ErrorResponse'

  /cleanse/issues/{issueId}/history:
    get:
      tags: [Cleanse]
      operationId: getIssueHistory
      summary: Get issue history
      description: Returns paginated history/lineage entries for a specific issue.
      parameters:
        - name: issueId
          in: path
          required: true
          schema:
            type: string
        - $ref: '#/components/parameters/Skip'
        - name: top
          in: query
          schema:
            type: integer
            default: 50
            minimum: 1
            maximum: 100
      responses:
        '200':
          description: Issue history
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/IssueHistoryResponse'
        '400':
          description: Invalid pagination parameters
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/ErrorResponse'
        '499':
          description: Client closed request
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/ErrorResponse'

  # ═══════════════════════════════════════════════════════════════════
  # Cleanse Export — Ad-hoc full export
  # ═══════════════════════════════════════════════════════════════════
  /cleanse-export/start:
    post:
      tags: [Cleanse Export]
      operationId: startFullExport
      summary: Start an ad-hoc full export of all active issues
      description: |
        The export runs in a long-running background task with a distributed lock
        to prevent concurrent exports. Progress can be tracked via the returned
        export ID. Unlike the export phase of a cleanse analysis (which is
        incremental), this always exports all active issues. It does NOT update
        the incremental-export timestamp and does NOT send email notifications.
      responses:
        '202':
          description: Export started successfully
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/StartExportResponse'
        '409':
          description: An export is already running
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/ExportErrorResponse'

  /cleanse-export/{exportId}:
    get:
      tags: [Cleanse Export]
      operationId: getExportOperation
      summary: Get progress and details of a specific export operation
      description: |
        Poll this endpoint to track export progress. When the export completes
        successfully, the response includes reportUrl (a presigned S3 download
        URL) and reportObjectKey. If the export fails, the error field
        contains the failure details.
      parameters:
        - name: exportId
          in: path
          required: true
          description: The export operation ID returned from the start endpoint
          schema:
            type: string
      responses:
        '200':
          description: Export operation details
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/CleanseExportOperationDto'
        '404':
          description: Export operation not found
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/ExportErrorResponse'

  /cleanse-export/{exportId}/regenerate-url:
    post:
      tags: [Cleanse Export]
      operationId: regenerateExportUrl
      summary: Regenerate the presigned URL for an export report
      description: |
        Use this when the original presigned URL has expired. The S3 object key
        must exist (i.e., the export must have completed successfully). The newly
        generated URL is also persisted on the export operation document.
      parameters:
        - name: exportId
          in: path
          required: true
          description: The export operation ID
          schema:
            type: string
      responses:
        '200':
          description: URL regenerated successfully
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/RegenerateExportUrlResponse'
        '400':
          description: Export has no report file (may not have completed successfully)
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/ExportErrorResponse'
        '404':
          description: Export operation not found
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/ExportErrorResponse'

  /cleanse-export:
    get:
      tags: [Cleanse Export]
      operationId: getExportOperations
      summary: Get a paginated list of export operations
      description: Returns export operations in reverse chronological order (most recent first).
      parameters:
        - $ref: '#/components/parameters/Skip'
        - $ref: '#/components/parameters/Top'
      responses:
        '200':
          description: Paginated list of export operations
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/ExportOperationsResponse'
        '400':
          description: Invalid parameters
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/ExportErrorResponse'

  # ═══════════════════════════════════════════════════════════════════
  # Holdings
  # ═══════════════════════════════════════════════════════════════════
  /holdings/cts/{lidFullIdentifier}:
    get:
      tags: [Holdings]
      operationId: getCtsCphHolding
      summary: Get a CTS CPH holding by LID
      description: "Looks up a CTS CPH holding by its LID full identifier (format: XX-CC/PPP/HHHH)."
      parameters:
        - name: lidFullIdentifier
          in: path
          required: true
          description: "LID full identifier (e.g., 'AB-12/345/6789')"
          schema:
            type: string
      responses:
        '200':
          description: CTS holding details
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/CtsCphHoldingResponse'
        '400':
          description: Invalid LID format
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/ErrorResponse'
        '404':
          description: Holding not found
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/ErrorResponse'
        '499':
          description: Client closed request
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/ErrorResponse'

  /holdings/sam/{cph}:
    get:
      tags: [Holdings]
      operationId: getSamCphHolding
      summary: Get a SAM CPH holding
      description: "Looks up a SAM CPH holding by its CPH value (format: CC/PPP/HHHH)."
      parameters:
        - name: cph
          in: path
          required: true
          description: "CPH value (e.g., '12/345/6789')"
          schema:
            type: string
      responses:
        '200':
          description: SAM holding details
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/SamCphHoldingResponse'
        '400':
          description: Invalid CPH format
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/ErrorResponse'
        '404':
          description: Holding not found
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/ErrorResponse'
        '499':
          description: Client closed request
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/ErrorResponse'

  # ═══════════════════════════════════════════════════════════════════
  # Throttle Policies
  # ═══════════════════════════════════════════════════════════════════
  /throttle-policies:
    get:
      tags: [Throttle Policies]
      operationId: getAllThrottlePolicies
      summary: Get all throttle policies
      responses:
        '200':
          description: List of throttle policies
          content:
            application/json:
              schema:
                type: array
                items:
                  $ref: '#/components/schemas/ThrottlePolicy'
    post:
      tags: [Throttle Policies]
      operationId: createThrottlePolicy
      summary: Create a throttle policy
      requestBody:
        required: true
        content:
          application/json:
            schema:
              $ref: '#/components/schemas/CreateThrottlePolicyCommand'
      responses:
        '201':
          description: Policy created
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/ThrottlePolicy'
        '400':
          description: Validation error
        '409':
          description: Duplicate slug

  /throttle-policies/active:
    get:
      tags: [Throttle Policies]
      operationId: getActiveThrottlePolicy
      summary: Get the active throttle policy
      responses:
        '200':
          description: Active policy (or Normal fallback)
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/ThrottlePolicy'

  /throttle-policies/{slug}:
    get:
      tags: [Throttle Policies]
      operationId: getThrottlePolicyBySlug
      summary: Get a throttle policy by slug
      parameters:
        - name: slug
          in: path
          required: true
          schema:
            type: string
      responses:
        '200':
          description: Throttle policy
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/ThrottlePolicy'
        '404':
          description: Policy not found
    put:
      tags: [Throttle Policies]
      operationId: updateThrottlePolicy
      summary: Update a throttle policy
      parameters:
        - name: slug
          in: path
          required: true
          schema:
            type: string
      requestBody:
        required: true
        content:
          application/json:
            schema:
              $ref: '#/components/schemas/UpdateThrottlePolicyCommand'
      responses:
        '200':
          description: Policy updated
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/ThrottlePolicy'
        '400':
          description: Validation error
        '404':
          description: Policy not found
    delete:
      tags: [Throttle Policies]
      operationId: deleteThrottlePolicy
      summary: Delete a throttle policy
      parameters:
        - name: slug
          in: path
          required: true
          schema:
            type: string
      responses:
        '204':
          description: Policy deleted
        '400':
          description: Cannot delete (e.g., Normal policy)
        '404':
          description: Policy not found
        '409':
          description: Cannot delete active policy

  /throttle-policies/{slug}/activate:
    post:
      tags: [Throttle Policies]
      operationId: activateThrottlePolicy
      summary: Activate a throttle policy
      parameters:
        - name: slug
          in: path
          required: true
          schema:
            type: string
      responses:
        '200':
          description: Policy activated
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/ThrottlePolicy'
        '400':
          description: Cannot activate (e.g., Normal policy)
        '404':
          description: Policy not found

  /throttle-policies/deactivate:
    post:
      tags: [Throttle Policies]
      operationId: deactivateAllThrottlePolicies
      summary: Deactivate all policies
      description: Reverts to Normal fallback policy.
      responses:
        '204':
          description: All policies deactivated

  # ═══════════════════════════════════════════════════════════════════
  # External Catalogue
  # ═══════════════════════════════════════════════════════════════════
  /externalcatalogue/files:
    get:
      tags: [External Catalogue]
      operationId: getFilesReport
      summary: Get file catalogue report
      description: Returns a plain text report of available files for the specified source type.
      parameters:
        - name: sourceType
          in: query
          required: true
          schema:
            type: string
            enum: [internal, external]
        - name: days
          in: query
          required: true
          description: Number of days to look back
          schema:
            type: integer
      responses:
        '200':
          description: Plain text file report
          content:
            text/plain:
              schema:
                type: string
        '400':
          description: Invalid parameters

  /externalcatalogue/upload:
    post:
      tags: [External Catalogue]
      operationId: uploadFile
      summary: Upload a CSV file
      description: Uploads a file to internal S3 storage. Filename must match a dataset definition pattern.
      parameters:
        - name: objectKey
          in: query
          required: true
          description: Filename (object key) — no path separators
          schema:
            type: string
      requestBody:
        required: true
        content:
          multipart/form-data:
            schema:
              type: object
              properties:
                file:
                  type: string
                  format: binary
      responses:
        '200':
          description: File uploaded
          content:
            application/json:
              schema:
                type: object
                properties:
                  message:
                    type: string
                  objectKey:
                    type: string
                  size:
                    type: integer
                  contentType:
                    type: string
        '422':
          description: Validation error

  /externalcatalogue/upload-raw:
    post:
      tags: [External Catalogue]
      operationId: uploadRawFile
      summary: Upload raw CSV content
      description: Uploads raw file content to internal S3 storage. Alternative for testing.
      parameters:
        - name: objectKey
          in: query
          required: true
          schema:
            type: string
      requestBody:
        required: true
        content:
          text/csv:
            schema:
              type: string
              format: binary
          application/csv:
            schema:
              type: string
              format: binary
          text/plain:
            schema:
              type: string
              format: binary
      responses:
        '200':
          description: File uploaded
          content:
            application/json:
              schema:
                type: object
                properties:
                  message:
                    type: string
                  objectKey:
                    type: string
                  size:
                    type: integer
                  contentType:
                    type: string
        '422':
          description: Validation error

  # ═══════════════════════════════════════════════════════════════════
  # Benchmark
  # ═══════════════════════════════════════════════════════════════════
  /benchmark/start:
    post:
      tags: [Benchmark]
      operationId: startBenchmark
      summary: Start a benchmark run
      description: |
        Starts a benchmark run with the supplied (or default) configuration.
        The run executes in the background; poll GET /api/benchmark/status
        and GET /api/benchmark/report for progress and results.
        Creates temporary `_benchmark_*` collections that are cleaned up
        automatically on completion or cancellation.
      requestBody:
        required: false
        content:
          application/json:
            schema:
              $ref: '#/components/schemas/BenchmarkConfig'
      responses:
        '202':
          description: Benchmark started
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/StartBenchmarkResponse'
        '409':
          description: A benchmark is already running
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/ErrorResponse'

  /benchmark/cancel:
    post:
      tags: [Benchmark]
      operationId: cancelBenchmark
      summary: Cancel the running benchmark
      description: |
        Requests cancellation of the currently running benchmark.
        Temporary collections are cleaned up automatically.
      responses:
        '200':
          description: Cancellation requested
          content:
            application/json:
              schema:
                type: object
                properties:
                  message:
                    type: string
        '404':
          description: No benchmark is currently running
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/ErrorResponse'

  /benchmark/report:
    get:
      tags: [Benchmark]
      operationId: getBenchmarkReport
      summary: Get the last benchmark report
      description: |
        Returns the full benchmark report from the most recent completed
        (or cancelled) run. Includes scenario results, driver metrics,
        explain plans, dataset/index fingerprints, and noisy-neighbour
        analysis with severity levels, remediation guidance, and a
        cross-correlated probable cause.
      responses:
        '200':
          description: Benchmark report
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/BenchmarkReport'
        '404':
          description: No benchmark report available
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/ErrorResponse'

  /benchmark/status:
    get:
      tags: [Benchmark]
      operationId: getBenchmarkStatus
      summary: Get benchmark subsystem status
      description: Returns whether a benchmark is running and summary info about the last report.
      responses:
        '200':
          description: Benchmark status
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/BenchmarkStatus'

# ═════════════════════════════════════════════════════════════════════
# Components
# ═════════════════════════════════════════════════════════════════════
components:
  parameters:
    Skip:
      name: skip
      in: query
      schema:
        type: integer
        default: 0
        minimum: 0
    Top:
      name: top
      in: query
      schema:
        type: integer
        default: 10
        minimum: 1
        maximum: 100

  schemas:
    # ── Shared ────────────────────────────────────────────────────
    ErrorResponse:
      type: object
      required: [message]
      properties:
        message:
          type: string
        timestamp:
          type: string
          format: date-time

    # ── Import ────────────────────────────────────────────────────
    StartBulkImportResponse:
      type: object
      properties:
        importId:
          type: string
          format: uuid
        sourceType:
          type: string
        message:
          type: string
        startedAt:
          type: string
          format: date-time

    ImportSummariesResponse:
      type: object
      properties:
        skip:
          type: integer
        top:
          type: integer
        count:
          type: integer
        imports:
          type: array
          items:
            type: object
        timestamp:
          type: string
          format: date-time

    ImportReport:
      type: object
      description: Full import report (structure varies).

    FileReportsResponse:
      type: object
      properties:
        importId:
          type: string
          format: uuid
        totalFiles:
          type: integer
        files:
          type: array
          items:
            type: object
        timestamp:
          type: string
          format: date-time

    PaginatedLineageEvents:
      type: object
      properties:
        collectionName:
          type: string
        recordId:
          type: string
        skip:
          type: integer
        top:
          type: integer
        count:
          type: integer
        totalEvents:
          type: integer
        events:
          type: array
          items:
            type: object

    GenerateRecordIdRequest:
      type: object
      required: [keyParts]
      properties:
        keyParts:
          type: array
          items:
            type: string
          example: ["NORTH", "F001"]

    GenerateRecordIdResponse:
      type: object
      properties:
        recordId:
          type: string
          description: URL-safe SHA256 hash (43 characters)
        keyParts:
          type: array
          items:
            type: string
        timestamp:
          type: string
          format: date-time

    DeleteCollectionResponse:
      type: object
      properties:
        collectionName:
          type: string
        success:
          type: boolean
        message:
          type: string
        deletedAtUtc:
          type: string
          format: date-time

    DeleteCollectionsResponse:
      type: object
      properties:
        deletedCollections:
          type: array
          items:
            type: string
        totalCount:
          type: integer
        success:
          type: boolean
        message:
          type: string
        deletedAtUtc:
          type: string
          format: date-time

    ClearDownStorageResponse:
      type: object
      properties:
        deletedKeys:
          type: array
          items:
            type: string
        totalDeleted:
          type: integer
        success:
          type: boolean
        message:
          type: string
        deletedAtUtc:
          type: string
          format: date-time

    # ── Query ─────────────────────────────────────────────────────
    QueryResult:
      type: object
      properties:
        collectionName:
          type: string
        count:
          type: integer
        totalCount:
          type: integer
          nullable: true
        data:
          type: array
          items:
            type: object
        executedAtUtc:
          type: string
          format: date-time

    # ── Cleanse — Operations ──────────────────────────────────────
    StartAnalysisResponse:
      type: object
      properties:
        operationId:
          type: string
        status:
          type: string
        message:
          type: string
        startedAtUtc:
          type: string
          format: date-time

    CancelAnalysisResponse:
      type: object
      properties:
        message:
          type: string
        requestedAtUtc:
          type: string
          format: date-time

    AnalysisRunsResponse:
      type: object
      properties:
        skip:
          type: integer
        top:
          type: integer
        count:
          type: integer
        runs:
          type: array
          items:
            $ref: '#/components/schemas/CleanseAnalysisOperationSummaryDto'
        timestamp:
          type: string
          format: date-time

    CleanseAnalysisOperationDto:
      type: object
      properties:
        id:
          type: string
        status:
          type: string
          enum: [Running, Completed, Failed, Cancelling, Cancelled]
        startedAtUtc:
          type: string
          format: date-time
        completedAtUtc:
          type: string
          format: date-time
          nullable: true
        progressPercentage:
          type: number
          format: double
          description: Weighted aggregate across all phases (0-100)
        statusDescription:
          type: string
        recordsAnalyzed:
          type: integer
        totalRecords:
          type: integer
        issuesFound:
          type: integer
        issuesResolved:
          type: integer
        error:
          type: string
          nullable: true
        durationMs:
          type: integer
          format: int64
          nullable: true
        reportObjectKey:
          type: string
          nullable: true
        reportUrl:
          type: string
          nullable: true
        finalAverageRpm:
          type: number
          format: double
          nullable: true
        cancelledAtUtc:
          type: string
          format: date-time
          nullable: true
        progress:
          nullable: true
          description: Unified operation tree progress snapshot. Null until the first periodic flush.
          $ref: '#/components/schemas/OperationNode'

    CleanseAnalysisOperationSummaryDto:
      type: object
      properties:
        id:
          type: string
        status:
          type: string
        startedAtUtc:
          type: string
          format: date-time
        completedAtUtc:
          type: string
          format: date-time
          nullable: true
        progressPercentage:
          type: number
          format: double
        recordsAnalyzed:
          type: integer
        totalRecords:
          type: integer
        issuesFound:
          type: integer
        issuesResolved:
          type: integer
        durationMs:
          type: integer
          format: int64
          nullable: true
        reportObjectKey:
          type: string
          nullable: true
        reportUrl:
          type: string
          nullable: true
        finalAverageRpm:
          type: number
          format: double
          nullable: true
        cancelledAtUtc:
          type: string
          format: date-time
          nullable: true

    RegenerateUrlResponse:
      type: object
      properties:
        operationId:
          type: string
        objectKey:
          type: string
        reportUrl:
          type: string
        regeneratedAtUtc:
          type: string
          format: date-time

    DeleteDataResponse:
      type: object
      properties:
        success:
          type: boolean
        collectionName:
          type: string
        deletedCount:
          type: integer
          format: int64
        message:
          type: string
        deletedAtUtc:
          type: string
          format: date-time

    TestNotificationResponse:
      type: object
      properties:
        success:
          type: boolean
        recipient:
          type: string
        notificationId:
          type: string
          nullable: true
        message:
          type: string
        sentAtUtc:
          type: string
          format: date-time

    # ── Cleanse — Issues ──────────────────────────────────────────
    IssuesResponse:
      type: object
      properties:
        skip:
          type: integer
        top:
          type: integer
        count:
          type: integer
        totalCount:
          type: integer
        issues:
          type: array
          items:
            $ref: '#/components/schemas/IssueDto'
        timestamp:
          type: string
          format: date-time

    IssueDto:
      type: object
      properties:
        id:
          type: string
        issueCode:
          type: string
        ruleCode:
          type: string
        errorCode:
          type: string
        cph:
          type: string
        ctsLidFullIdentifier:
          type: string
          nullable: true
        description:
          type: string
        isActive:
          type: boolean
        isIgnored:
          type: boolean
        resolutionStatus:
          type: string
        assignedTo:
          type: string
          nullable: true
        operationId:
          type: string
        createdAtUtc:
          type: string
          format: date-time
        lastUpdatedAtUtc:
          type: string
          format: date-time

    IssueCommandResponse:
      type: object
      properties:
        issueId:
          type: string
        message:
          type: string
        timestamp:
          type: string
          format: date-time

    IssueHistoryResponse:
      type: object
      properties:
        issueId:
          type: string
        skip:
          type: integer
        top:
          type: integer
        count:
          type: integer
        entries:
          type: array
          items:
            $ref: '#/components/schemas/IssueHistoryEntryDto'
        timestamp:
          type: string
          format: date-time

    IssueHistoryEntryDto:
      type: object
      properties:
        id:
          type: string
        action:
          type: string
        performedBy:
          type: string
          nullable: true
        detail:
          type: string
          nullable: true
        timestamp:
          type: string
          format: date-time

    PerformedByRequest:
      type: object
      required: [performedBy]
      properties:
        performedBy:
          type: string

    UpdateResolutionStatusRequest:
      type: object
      required: [status, performedBy]
      properties:
        status:
          type: string
          description: "None, Todo, InProgress, Resolved"
          enum: [None, Todo, InProgress, Resolved]
        performedBy:
          type: string

    AssignIssueRequest:
      type: object
      required: [assignedTo, performedBy]
      properties:
        assignedTo:
          type: string
        performedBy:
          type: string

    # ── Cleanse Export ────────────────────────────────────────────
    StartExportResponse:
      type: object
      required: [exportId, status, message]
      properties:
        exportId:
          type: string
          description: Unique identifier for the started export operation. Use this to poll progress.
        status:
          type: string
          description: Initial status of the export (Pending).
        message:
          type: string
        startedAtUtc:
          type: string
          format: date-time

    CleanseExportOperationDto:
      type: object
      description: Full details of an ad-hoc cleanse export operation.
      properties:
        id:
          type: string
          description: Unique export operation identifier.
        status:
          type: string
          enum: [Pending, Running, Completed, Failed]
        startedAtUtc:
          type: string
          format: date-time
        completedAtUtc:
          type: string
          format: date-time
          nullable: true
        progressPercentage:
          type: number
          format: double
          description: Progress percentage (0-100).
        statusDescription:
          type: string
          description: Human-readable status description.
        totalRecords:
          type: integer
          description: Total number of records to export.
        recordsExported:
          type: integer
          description: Number of records exported so far.
        reportObjectKey:
          type: string
          nullable: true
          description: S3 object key for the generated report. Populated on successful completion.
        reportUrl:
          type: string
          nullable: true
          description: Presigned URL to download the report. Use regenerate-url if expired.
        error:
          type: string
          nullable: true
          description: Error message with exception details if the export failed.
        durationMs:
          type: integer
          format: int64
          nullable: true
          description: Total duration in milliseconds when complete.

    CleanseExportOperationSummaryDto:
      type: object
      description: Lightweight summary of an ad-hoc export operation for listing.
      properties:
        id:
          type: string
        status:
          type: string
          enum: [Pending, Running, Completed, Failed]
        startedAtUtc:
          type: string
          format: date-time
        completedAtUtc:
          type: string
          format: date-time
          nullable: true
        progressPercentage:
          type: number
          format: double
        totalRecords:
          type: integer
        recordsExported:
          type: integer
        reportObjectKey:
          type: string
          nullable: true
        reportUrl:
          type: string
          nullable: true
        durationMs:
          type: integer
          format: int64
          nullable: true

    RegenerateExportUrlResponse:
      type: object
      required: [exportId, objectKey, reportUrl]
      properties:
        exportId:
          type: string
          description: The export operation ID.
        objectKey:
          type: string
          description: The S3 object key.
        reportUrl:
          type: string
          description: The new presigned URL.
        regeneratedAtUtc:
          type: string
          format: date-time

    ExportOperationsResponse:
      type: object
      required: [exports]
      properties:
        skip:
          type: integer
        top:
          type: integer
        count:
          type: integer
          description: Actual number of export operations returned.
        exports:
          type: array
          items:
            $ref: '#/components/schemas/CleanseExportOperationSummaryDto'
        timestamp:
          type: string
          format: date-time

    ExportErrorResponse:
      type: object
      required: [message]
      description: Error response for export operations.
      properties:
        message:
          type: string
          description: The error message with details.
        timestamp:
          type: string
          format: date-time

    # ── Holdings ──────────────────────────────────────────────────
    CtsCphHoldingResponse:
      type: object
      properties:
        lidFullIdentifier:
          type: string
        locationName:
          type: string
          nullable: true
        holding:
          type: object
          additionalProperties: true
        keepers:
          $ref: '#/components/schemas/QueryResult'
        timestamp:
          type: string
          format: date-time

    SamCphHoldingResponse:
      type: object
      properties:
        cph:
          type: string
        locationName:
          type: string
          nullable: true
        holding:
          type: object
          additionalProperties: true
        herd:
          $ref: '#/components/schemas/QueryResult'
        parties:
          $ref: '#/components/schemas/QueryResult'
        holders:
          $ref: '#/components/schemas/QueryResult'
        timestamp:
          type: string
          format: date-time

    # ── Throttle Policies ─────────────────────────────────────────
    ThrottlePolicy:
      type: object
      properties:
        name:
          type: string
        slug:
          type: string
        isActive:
          type: boolean
        isReadOnly:
          type: boolean
        settings:
          $ref: '#/components/schemas/ThrottlePolicySettings'

    ThrottlePolicySettings:
      type: object
      properties:
        ingestion:
          type: object
        cleanseAnalysis:
          type: object
          properties:
            pumpBatchSize:
              type: integer
            pumpDelayMs:
              type: integer
            recordIssueDelayMs:
              type: integer
            progressUpdateInterval:
              type: integer
            rpmWindowSeconds:
              type: integer
        cleanseExport:
          type: object
          properties:
            streamBatchSize:
              type: integer
            throttlingDelayMs:
              type: integer
            rpmWindowSeconds:
              type: integer
        issueDeactivation:
          type: object
          properties:
            batchSize:
              type: integer
            throttleDelayMs:
              type: integer
            rpmWindowSeconds:
              type: integer
        issueQuery:
          type: object

    CreateThrottlePolicyCommand:
      type: object
      required: [name]
      properties:
        name:
          type: string
        settings:
          $ref: '#/components/schemas/ThrottlePolicySettings'

    UpdateThrottlePolicyCommand:
      type: object
      properties:
        name:
          type: string
        settings:
          $ref: '#/components/schemas/ThrottlePolicySettings'

    # ── Benchmark ────────────────────────────────────────────────
    BenchmarkConfig:
      type: object
      description: |
        Configuration for a benchmark run. All values have safe defaults
        designed to avoid overloading a shared production database.
        Omit the body entirely to use all defaults.
      properties:
        seedCount:
          type: integer
          default: 10000
          description: Number of deterministic seed records to create.
        concurrency:
          type: integer
          default: 4
          description: Maximum degree of parallelism for scenario execution.
        duration:
          type: string
          default: "00:03:00"
          description: "Total benchmark duration (TimeSpan format, e.g. '00:03:00'). Scenarios loop until this elapses."
        throttleDelay:
          type: string
          default: "00:00:00.0100000"
          description: "Delay injected between every Mongo operation (TimeSpan format, e.g. '00:00:00.010')."
        collectionPrefix:
          type: string
          default: "_benchmark_"
          description: Prefix applied to all temporary benchmark collections.

    StartBenchmarkResponse:
      type: object
      properties:
        message:
          type: string
        config:
          $ref: '#/components/schemas/BenchmarkConfig'

    BenchmarkStatus:
      type: object
      properties:
        isRunning:
          type: boolean
        hasReport:
          type: boolean
        lastReportStatus:
          type: string
          nullable: true
          description: "Status of the last report: Completed, Cancelled, or Failed."
        lastReportTimestamp:
          type: string
          format: date-time
          nullable: true

    BenchmarkReport:
      type: object
      description: |
        Full benchmark report. Designed to be JSON-serialised, written to disk,
        and compared across environments to detect noisy-neighbour impact.
      properties:
        environment:
          type: string
          description: Machine name where the benchmark ran.
        timestampUtc:
          type: string
          format: date-time
        config:
          $ref: '#/components/schemas/BenchmarkConfig'
        status:
          type: string
          description: "Completed, Cancelled, or Failed: <message>."
        totalElapsedSeconds:
          type: number
          format: double
        datasetFingerprints:
          type: array
          items:
            $ref: '#/components/schemas/DatasetFingerprint'
        indexFingerprints:
          type: array
          items:
            $ref: '#/components/schemas/IndexFingerprint'
        scenarioResults:
          type: array
          description: "Results for each scenario: PointLookup, RangeQuery, Aggregation, BulkWrite, MiniETL."
          items:
            $ref: '#/components/schemas/ScenarioResult'
        driverMetrics:
          $ref: '#/components/schemas/DriverMetrics'
        explainResults:
          type: array
          items:
            $ref: '#/components/schemas/ExplainResult'
        noisyNeighbourAnalysis:
          $ref: '#/components/schemas/NoisyNeighbourAnalysis'

    DatasetFingerprint:
      type: object
      description: Fingerprint of a benchmark collection's data distribution.
      properties:
        collectionName:
          type: string
        documentCount:
          type: integer
          format: int64
        avgDocumentSizeBytes:
          type: number
          format: double
        p95DocumentSizeBytes:
          type: number
          format: double

    IndexFingerprint:
      type: object
      description: Fingerprint of an index on a benchmark collection.
      properties:
        collectionName:
          type: string
        indexName:
          type: string
        keyDefinition:
          type: object
          additionalProperties: true
          description: BSON key definition document.
        isUnique:
          type: boolean

    ScenarioResult:
      type: object
      description: Result of a single benchmark scenario.
      properties:
        scenarioName:
          type: string
        totalOperations:
          type: integer
        errorCount:
          type: integer
        elapsedSeconds:
          type: number
          format: double
        opsPerSecond:
          type: number
          format: double
          description: Wall-clock throughput.
        effectiveOpsPerSecond:
          type: number
          format: double
          description: Throughput excluding throttle wait time — pure Mongo performance.
        latency:
          $ref: '#/components/schemas/LatencyStats'

    LatencyStats:
      type: object
      description: Percentile latency statistics in milliseconds.
      properties:
        avgMs:
          type: number
          format: double
        p50Ms:
          type: number
          format: double
        p95Ms:
          type: number
          format: double
        p99Ms:
          type: number
          format: double
        minMs:
          type: number
          format: double
        maxMs:
          type: number
          format: double

    DriverMetrics:
      type: object
      description: Aggregated MongoDB driver command and connection-pool statistics.
      properties:
        commandLatency:
          type: object
          additionalProperties:
            $ref: '#/components/schemas/LatencyStats'
          description: "Per-command-type latency breakdown (e.g. find, update, aggregate)."
        commandFailures:
          type: object
          additionalProperties:
            type: integer
          description: Total count of failed commands by command name.
        connectionCheckoutWait:
          $ref: '#/components/schemas/LatencyStats'
          nullable: true
          description: Connection checkout wait-time statistics.
        checkoutFailures:
          type: integer
          description: Number of checkout failures (pool exhausted).
        connectionsCreated:
          type: integer
        connectionsClosed:
          type: integer
        poolClearedEvents:
          type: integer
          description: Number of times the driver was forced to reset the connection pool.

    ExplainResult:
      type: object
      description: Captured explain-plan output for a key query.
      properties:
        queryName:
          type: string
          description: "Name of the explained query: PointLookup, RangeQuery, or Aggregation."
        winningPlan:
          type: string
          description: JSON representation of the winning query plan.
        totalDocsExamined:
          type: integer
          format: int64
        totalKeysExamined:
          type: integer
          format: int64
        nReturned:
          type: integer
          format: int64
        rawExplain:
          type: object
          additionalProperties: true
          nullable: true
          description: Full raw explain output (BSON document).

    NoisyNeighbourAnalysis:
      type: object
      description: |
        Diagnostic analysis for noisy-neighbour indicators. Cross-correlates
        connection pool health, command latency, and scenario errors to
        produce a probable root cause.
      properties:
        hasRedFlags:
          type: boolean
          description: True if any red flag fired.
        overallRisk:
          type: string
          enum: [None, Warning, Critical]
          description: Highest severity across all flags.
        probableCause:
          type: string
          nullable: true
          description: |
            Plain-language root-cause diagnosis based on cross-correlation
            of multiple flags. Null when no flags are present.
        flags:
          type: array
          items:
            $ref: '#/components/schemas/RedFlag'

    RedFlag:
      type: object
      description: A single noisy-neighbour diagnostic flag.
      properties:
        category:
          type: string
          description: Machine-readable category for comparison tooling.
        severity:
          type: string
          enum: [None, Warning, Critical]
        description:
          type: string
          description: Human-readable description of the problem observed.
        remediation:
          type: string
          description: Actionable guidance on what to investigate next.
        observedValue:
          type: number
          format: double
          description: The observed value that triggered the flag.
        threshold:
          type: number
          format: double
          description: The threshold that was exceeded.

    # ── Timings ──────────────────────────────────────────────────
    OperationNode:
      type: object
      nullable: true
      description: |
        Immutable snapshot of a single node in the operation tree.
        Combines timing, progress, and rate metrics into one self-describing
        recursive structure. The root node is named "total" with phase children
        (Analysis, Deactivation, Export) that may contain further sub-scopes.
      properties:
        name:
          type: string
          description: "Node name (e.g. \"total\", \"Analysis\", \"CTS Pump\", \"fetching\")."
        status:
          type: string
          enum: [not-started, in-progress, completed, failed, cancelled]
          description: Current status of this node.
        description:
          type: string
          nullable: true
          description: Human-readable description of what this node is doing.
        percentComplete:
          type: number
          format: double
          nullable: true
          description: "Progress percentage (0-100). Leaf nodes compute from processedCount/totalRecords; parent nodes use weighted average of children."
        processedCount:
          type: integer
          nullable: true
          description: Number of records processed so far in this scope.
        totalRecords:
          type: integer
          nullable: true
          description: Total records expected in this scope.
        elapsedMs:
          type: integer
          format: int64
          description: Cumulative elapsed time in milliseconds.
        elapsed:
          type: string
          description: "Formatted elapsed time (e.g. \"00:07:42.3\")."
        projectedRemainingMs:
          type: integer
          format: int64
          nullable: true
          description: Estimated remaining time in milliseconds based on current throughput.
        projectedEndTimeUtc:
          type: string
          format: date-time
          nullable: true
          description: Projected completion time based on current throughput.
        currentRecordsPerMinute:
          type: number
          format: double
          nullable: true
          description: Windowed records-per-minute rate (last 60 seconds).
        averageRecordsPerMinute:
          type: number
          format: double
          nullable: true
          description: Average records-per-minute rate since the scope started.
        children:
          type: array
          nullable: true
          description: Child nodes forming a recursive tree.
          items:
            $ref: '#/components/schemas/OperationNode'

  securitySchemes:
    ApiKeyAuth:
      type: apiKey
      in: header
      name: X-API-Key
      description: API key authentication.

security:
  - ApiKeyAuth: []