What We Collect

Prizm extracts Sigma metadata via the Sigma REST API (/v2/...) and, for lineage and usage, the Sigma Logs API.

What this collects: Workspaces, Workbooks, Pages, Data Elements, and Datasets into the Prizm Asset Catalog.

Sigma object	Primary source	Key properties extracted
Workspace	`GET /v2/workspaces`	id, name, description, createdAt, updatedAt, url
Workbook	`GET /v2/workbooks`	id, name, description, createdAt, updatedAt, url, path, owner
Page	`GET /v2/workbooks/{id}/pages`	id, title, workbook_id
Data Element (Table)	`GET /v2/workbooks/{id}/elements`	displayName, type, columns, dataSourceId
Data Element (Pivot)	`GET /v2/workbooks/{id}/elements`	displayName, type, rowKeys, columnKeys, values
Data Element (Viz)	`GET /v2/workbooks/{id}/elements`	displayName, chartType, xAxis, yAxis, colorDimension
Data Element Field	`GET /v2/workbooks/{id}/elements/{eid}/columns`	displayName, formula, type (calculated or base)
Dataset	`GET /v2/datasets`	id, name, description, createdAt, updatedAt, url, warehouseSource
Tag	`GET /v2/workbooks/{id}/tags`	tag_name, tag_value, attached_asset_type, attached_asset_id

Metadata captured per Workbook: id, name, description, path, workspace_id, workspace_name, owner (display name and email), created_at, updated_at, deep-link URL back to Sigma, tag names/values, page count, and data element count.Metadata captured per Dataset: id, name, description, url, created_at, updated_at, and warehouse_source (connection name, database, schema, table) — used for lineage.

Pagination is handled transparently for Workbooks and Datasets with more than 100 items. Assets in workspaces excluded by the connector’s workspace filter are not catalogued. Endorsement level (certified / promoted) is not currently exposed by the Sigma API and is not captured.

Sigma’s semantic layer sits between raw warehouse tables and the end visualisation. Prizm distinguishes two field types:

Field type	Description and storage in Prizm
Base field	Direct reference to a warehouse column; stored with column name and warehouse path for lineage resolution
Calculated field	Field defined by a Sigma formula expression; stored with `displayName` and `sigmaDataElementFieldFormula`; marked as calculated in the catalog with a Calculated badge

Formula changes in calculated fields are detected as field-level schema drift — a formula diff between consecutive catalog runs generates a drift event in the Prizm observability feed.

What this collects: the full lineage chain from a warehouse table through a Sigma Dataset, Workbook, Page, and Data Element to end consumers.

Source (in priority order)	What it provides
Sigma Logs API (`GET /v2/usage/queries`)	Query logs recording which warehouse tables each Sigma element executed against — primary lineage source
Dataset metadata (`warehouseSource`)	Direct link from a Sigma Dataset to a warehouse table/schema
Data Element column references	Fallback resolution via base warehouse columns, using the warehouse connector already configured in Prizm

Asset	Upstream link	Downstream link
Dataset	Warehouse connection + database + schema + table	Workbooks and Data Elements referencing this Dataset
Workbook	Datasets used by its Data Elements	Pages; end consumers (usage log users)
Data Element	Parent Page + Workbook; upstream Dataset or direct warehouse columns	Visualisations or downstream consumers embedding this element
Calculated Field	Base fields or other calculated fields referenced in its formula	Other calculated fields that reference this field

Workbooks with no Dataset (querying the warehouse directly) still get lineage resolved via query log analysis. Lineage from an upstream dbt or ADF pipeline into a Sigma Dataset resolves automatically when both connectors are active in Prizm.

Criticality for Workbooks and Datasets uses the same BI-asset framework as Power BI and Tableau: 60% Dependency (lineage to underlying sources) + 40% Usage (unique users). Scores are re-evaluated whenever upstream table criticality changes or workbook usage changes materially.Monitoring signals:

Signal	Detection logic	Default severity
Zero views in 30 days	No usage log entries in a 30-day window	Low
Dataset not refreshed	Dataset `last_updated` age exceeds the refresh SLA	High
Upstream schema change	A warehouse column referenced by a Sigma Data Element is dropped or renamed	Critical
Calculated field formula change	Formula diff between consecutive Catalog Job runs	Medium
Usage spike / drop	30-day rolling view count deviates >50% from the 90-day baseline	Medium

Stale workbooks appear in the observability feed within one hour of an SLA breach (Operational job cycle). Upstream schema-change alerts propagate from the warehouse connector to affected Sigma Workbooks via lineage within one evaluation cycle. Zero-view workbooks are listed in the unused-assets report with last view date and owner.

Metric	Workbook	Dataset	Source
Freshness	Default enabled	Default enabled	`last_updated` from the Sigma API; age = now − last_updated
Reliability (refresh success rate)	Default enabled	Default enabled	Failed refresh events from usage log / Dataset refresh status
Usage (view count)	Default disabled	Default disabled	`view_count_30d` from the Sigma usage log
Unique users (30d)	Default disabled	Default disabled	Distinct users from the Sigma usage log
DQ Score	Propagated from upstream tables	Propagated from upstream tables	Inherited via lineage from the upstream warehouse table/dataset

Reliability and Usage metrics depend on Sigma’s audit/query log access, which requires a Premier-tier Sigma account with logging enabled. On accounts without that access, these metrics are unavailable rather than approximated.

DQ Score propagation is controlled by the Propagate Scores Based On toggle in the connector configuration and can be disabled per connection without affecting the upstream warehouse alerts themselves.

Sigma extends the same Prizm Alert Module framework used across all connectors:

Alert	Trigger	Propagation path	Severity
Upstream DQ rule violation	A quality rule fires on a warehouse table feeding a Sigma Dataset	Warehouse Table → Sigma Dataset → Workbook → Data Element	Inherited
Upstream schema change	Column referenced by a Sigma Data Element base field is dropped or renamed	Warehouse Table → Sigma Dataset → affected Data Elements	Critical
Calculated field formula change	Formula diff between catalog runs	Sigma Data Element Field	Medium
Workbook freshness breach	`last_updated` exceeds SLA threshold	Sigma Workbook	High
Unused workbook	Zero views in a 30-day window	Sigma Workbook	Low (governance)
Dataset refresh failure	`last_updated` stale beyond the configured window	Sigma Dataset	High

Propagated alerts carry the upstream source asset name and lineage path in the notification payload, and the DQ score shown on a Sigma asset indicates whether it’s native to Sigma or propagated from upstream.Event & workflow management: freshness breaches, refresh failures, upstream schema changes, and unused-workbook signals automatically create Prizm issues for Critical and High severity events. Each issue links to the affected Sigma asset with a deep-link back to the Workbook in Sigma, and Jira/ADO integration cascades fire from Sigma-sourced issues the same way they do for any other connector.

Get Started

Core Concepts

Platform

Architecture

Data Sources

Integrations

Security & Access Control

Deployment

Help

Settings

MCP

Release Notes

API Reference

Next Steps

Setup

FAQ

​Next Steps

Setup

FAQ

Next Steps