What Prizm Collects from Tableau

Prizm runs three platform jobs against each Tableau connector. Select a job below to see what it collects and how.

Catalog / Metadata
Operational / Refresh
Usage / Adoption


Type	CONTEXT
Execution	Daily; can also be triggered on-demand after a publish

What this job does: Full inventory of all Tableau objects using the Tableau Metadata API (GraphQL). This is the foundation job — it powers the Prizm catalog, lineage graph, and asset detail pages. If this job is incorrect, all downstream data is affected.

Object	Fields Collected	Source
Site	name, site ID	Tableau Metadata API
Project	name, description, owner, projectHierarchy, isTopLevelProject, sourceCreatedAt, sourceUpdatedAt	Tableau Metadata API
Workbook	name, description, sourceURL, sourceOwner, projectHierarchy, sourceCreatedAt, sourceUpdatedAt	Tableau Metadata API
Dashboard	name, sourceURL, projectHierarchy, sourceCreatedAt, sourceUpdatedAt	Tableau Metadata API
Worksheet	name, sourceURL, projectHierarchy, sourceCreatedAt, sourceUpdatedAt	Tableau Metadata API
Published Data Source	name, description, sourceOwner, isPublished, hasExtracts, upstreamTables, upstreamDatabases, certificateStatus, certifier, projectHierarchy	Tableau Metadata API
Datasource Field	name, description, upstreamTables, upstreamColumns, dataCategory, role (dimension/measure), dataType, formula, binSize, datasourceFieldType	Tableau Metadata API
Calculated Field	name, description, dataCategory, role, tableauDataType, formula	Tableau Metadata API

Tableau Metrics (native Tableau feature) are deprecated and are not ingested. Lineage captured here is logical lineage — based on the declared upstream tables in the datasource definition, not runtime query execution.


Type	TRUST
Execution	Hourly

What this job does: Monitors extract refresh health in near-real-time using the Tableau REST API. Answers “Is data fresh?”, “Are refreshes failing?”, and “Is something broken right now?” Only tracks extract refresh tasks and refresh-related background jobs — permission jobs, subscription sends, and thumbnail generation are excluded.

Signal	Source	Description
Freshness delay	`last_successful_refresh_time` via Extract Refresh Tasks API	Time elapsed since last successful refresh. Dynamic threshold: expected refresh interval ± 3× average execution time.
Refresh duration	Background jobs API	Time taken for each extract refresh to complete
Refresh success rate	Background jobs API	% of refresh jobs completing successfully
Consecutive failures	Background jobs API	Number of consecutive failed refresh jobs on a datasource

This job runs hourly so freshness SLA breaches surface quickly. The anomaly detection threshold is auto-calculated from the configured refresh schedule and historical execution times — no manual threshold setup is required.

Signals this job powers:

freshness_delay_minutes
refresh_success_rate
consecutive_failures
p95_refresh_duration


Type	CONTEXT
Execution	Daily

What this job does: Collects workbook view counts and user activity data to answer “Who cares if this breaks?”, “What is safe to deprecate?”, and “What should be prioritized?”. Uses the Tableau Cloud Usage Stats API (Tableau Cloud) or Admin Views / Repository (Tableau Server).

Metric	Source	Description
hits_last_14d	`hitsLastTwoWeeksTotal`	Views in the last 14 days
hits_last_30d	`hitsLastOneMonthTotal`	Views in the last 30 days
hits_last_90d	`hitsLastThreeMonthsTotal`	Views in the last 90 days
hits_last_365d	`hitsLastTwelveMonthsTotal`	Views in the last 365 days
hits_total	`hitsTotal`	All-time view count
daily_active_users	Admin Insights	Distinct users per day
usage_weighted_health_score	Computed	Quality score weighted by usage — surfaces high-traffic workbooks with issues first

Usage data feeds zombie dashboard detection — workbooks with no views in 90+ days are flagged for deprecation review.

Tableau Cloud requires API version 3.17 or later for usage stats. For Tableau Server, Admin Views access requires a user with repository read access.

What You See in Prizm

Once Prizm completes its initial catalog run, every connected Tableau workbook gets a unified asset detail page. Worksheets and dashboards are accessible from the workbook detail page.

Overview
Details
Attribute
Lineage

Section	What Prizm Shows
Quality Score	Average quality score across all attributes of the mapped upstream assets
Alerts	Active alerts propagated from upstream tables
Issues	Open issues linked to this workbook
Views	Total view count and recent activity
Last Updated	Timestamp of last metadata refresh from Tableau
Score Trend	Time-series chart of quality score, alerts, and issues over selected dates
Semantic Context	Domain, Application, Product, Tag, and Terms classifications
Owners	Tableau asset owner synced from Tableau metadata

The score trend chart shows how data quality observations for impacted upstream tables change over time. Score is shown as green bars; alerts as a blue line; issues as a red line.Below the chart, a list of associated upstream assets is shown with columns: Account, Alerts, Application, Asset Type, Connection Name, Created On, Description, Domain, Issues, Last Run Date, Product, Score, Tag, Term.

The Details tab lists all worksheets and dashboards contained in the workbook.

Column	Description
Name	Worksheet or dashboard name (click to open detail view)
Type	Worksheet or Dashboard
Site	Tableau site
Owner	Asset owner
Columns	Number of fields used
Tables	Number of upstream tables referenced
Score	Propagated quality score
Alerts	Active alert count
Issues	Open issue count
Created Date	Asset creation timestamp
Updated Date	Last updated timestamp

Click any row to open the individual worksheet or dashboard detail view, which includes its own Overview, Attributes, and Lineage tabs plus a preview of the sheet or dashboard in the metadata pane.

The Attribute tab lists all fields used across the workbook.

Column	Description
Name	Field name
Data Type	Tableau data type
Connection Type	Source connection (e.g., Snowflake, Databricks)
Database	Upstream database
Schema	Upstream schema
Table	Upstream table
Definition	Field formula or expression (for calculated fields)
Description	Field description
Is Null	Whether the field allows null values
Score	Quality score from the upstream column
Semantics	Semantic classification (domain, term)
Sheet	Worksheet(s) where this field is used

The Lineage tab renders a directed graph showing the full data flow from upstream tables through Tableau layers.

Layer	What Prizm Shows
Data Connections	Source database tables and views
Published Data Sources	Tableau datasources that transform or combine upstream data
Worksheets	Individual sheets consuming the datasource
Dashboards	Dashboards composed from one or more worksheets

Lineage is auto-inferred from the upstreamTables and upstreamColumns declared in the Tableau datasource definition. It can also be manually mapped by selecting asset names from the dropdown on the Lineage tab. Attribute-level (column-to-field) lineage is visible at each layer.

Metadata Pane

Every Tableau asset includes a metadata pane with the following information pulled directly from Tableau:

Field	Description
Description	Pulled from Tableau metadata if available; otherwise AI-generated
Tags	Tags synced from Tableau
Owner Name	User who owns the Tableau asset
Site	Tableau site where the asset is published
Updated At	Date and time of last update in Tableau
# of Tables	Total upstream tables used in the asset
# of Attributes	Total fields/columns used in the asset
# of Sheets	Number of worksheets in the workbook
# of Dashboards	Number of dashboards in the workbook

The metadata pane also shows semantic layer classifications (Domains, Tags, Products, Application, Terms) and upstream/downstream system links.

Get Started

Core Concepts

Platform

Architecture

Data Sources

Integrations

Security & Access Control

Deployment

Help

Settings

MCP

Release Notes

API Reference

What Prizm Collects from Tableau

What You See in Prizm

Metadata Pane

Next Steps

Setup

FAQ

​What You See in Prizm

​Metadata Pane

​Next Steps

Setup

FAQ

What You See in Prizm

Metadata Pane

Next Steps