Query Admin Insights — Job Performance

Issues a VizQL Data Service (VDS) query against the Admin Insights Job Performance published datasource on the connected Tableau Cloud site. Returns records of extract refresh jobs, subscription jobs, flow runs, and bridge jobs.

The tool is admin-only — it is registered only when ADMIN_TOOLS_ENABLED=true, and at request time it verifies the caller's site role and rejects anything below SiteAdministratorCreator / SiteAdministratorExplorer / ServerAdministrator. The Admin Insights datasource LUID is resolved automatically; callers do not pass datasourceLuid.

APIs called

• Query Datasource (VDS)
• Query Data Sources (REST) — used internally to resolve the Job Performance dataset LUID
• Get User on Site (REST) — used internally for the admin gate

Required arguments

query

A fully formed VDS query object: fields, filters, parameters. The schema mirrors the schema accepted by the query-datasource tool.

The fields array must contain at least one entry — omitting it causes a VDS error.

Job Type values are stored without spaces, e.g. RefreshExtracts, IncrementExtracts, RefreshExtractsViaBridge, IncrementExtractsViaBridge, SendSingleSubscription, RunFlow. Query the Job Type field alone to list the values present on a site.

Common usage:

• Analyze extract refresh durations and failure rates per datasource or workbook.
• Identify extracts with high consecutive failure counts or long run times.
• Compare scheduled frequency against actual job completion times to recommend schedule optimization.
• Find jobs that overlap or compete for resources in the same time window.

Available fields

Category	Fields
Job identity	Job ID, Job LUID, Job Type, Job Result, Final Job Result
Item details	Item ID, Item LUID, Item Name, Item Type, Item Hyperlink
Timing (UTC)	Created At, Queued At, Started At, Completed At, Overflow Queued At
Timing (local)	Created At (local), Queued At (local), Started At (local), Completed At (local), Overflow Queued At (local)
Durations (seconds)	Job Duration, Job Execution Duration, Job Queued Duration, Job Overflow Queued Duration
Schedule	Schedule Name, Schedule LUID
Owner/Project	Owner Email, Parent Project Name, Parent Project Owner Email
Extract	Extract File Size
Subscription	Subscriber Email, Subscriber ID, Subscription Subject
Bridge	Agent Name, Agent Version, Agent Timezone, Agent is Pooled?, Pool Name, Bridge Started At, Bridge Completed At, Bridge Started At (Local), Bridge Completed At (Local), Bridge Refresh Duration, Bridge Extract Upload Duration, Bridge Job Result, Bridge Error Message, Bridge Error Type, Bridge Initiator User Name
Flags	Was Manual Run, Was Overflow Queued
Other	Error Message, Admin Insights Published At

Example:

{
  "query": {
    "fields": [
      { "fieldCaption": "Item Name" },
      { "fieldCaption": "Job Type" },
      { "fieldCaption": "Job Result" },
      { "fieldCaption": "Started At" },
      { "fieldCaption": "Job Duration" },
      { "fieldCaption": "Schedule Name" }
    ],
    "filters": [
      {
        "field": { "fieldCaption": "Job Type" },
        "filterType": "SET",
        "values": ["RefreshExtracts"],
        "exclude": false
      }
    ]
  }
}

Optional arguments

limit

The maximum number of rows to return. The tool will pass this through as rowLimit on the VDS request and additionally truncate the response if VDS returns more rows than requested.

Example: 500

See also: MAX_RESULT_LIMIT

Parameters

The Job Performance datasource exposes a Timezone parameter (integer offset, e.g. -7 for PDT). Local-time fields (Started At (local), Completed At (local), etc.) are adjusted by this parameter. Pass it via the parameters field of the query object if you need local-time output.

Notes and caveats

• Tableau Cloud Job Performance lookback caps at 90 days by default (365 days with Advanced Management).
• The datasource does not support server-side pagination — use filters and limit to control response size. For large sites, always filter by Job Type and/or a date range on Started At.
• This tool intentionally bypasses the standard datasource access checker because the Admin Insights datasources are internal/known and admin-gated independently.
• For the complete field schema with data types and descriptions, use get-datasource-metadata with the Job Performance datasource LUID.

Example result

{
  "data": [
    {
      "Item Name": "Sales Pipeline",
      "Job Type": "RefreshExtracts",
      "Job Result": "Succeeded",
      "Started At": "2026-05-28T08:00:12Z",
      "Job Duration": 45.2,
      "Schedule Name": "Daily at 8am"
    },
    {
      "Item Name": "Marketing Dashboard",
      "Job Type": "RefreshExtracts",
      "Job Result": "Failed",
      "Started At": "2026-05-28T08:15:00Z",
      "Job Duration": 120.8,
      "Schedule Name": "Daily at 8am"
    }
  ]
}

Related

• The MCP prompt job-optimization-inform invokes this tool and renders the results as a job optimization report for human-in-the-loop review. It defaults to extract refresh jobs and can discover and analyze every job type on the site.