​

Features

This section contains the list of new features identified as part of the Sep 17 sprint (3.1.6).

​

New Oracle Database Connector

Prizm now connects to Oracle Database — on-premises, self-hosted VM, and Oracle Cloud Infrastructure (OCI), including Autonomous Database — on Oracle 19c and higher. Key Capabilities

• Full catalog extraction: schemas, tables, views, columns, stored procedures, and constraints, with descriptions pulled directly from Oracle’s native comments
• Authentication via username/password or enterprise vault integration (HashiCorp Vault, Azure Key Vault, AWS Secrets Manager), with TCPS/SSL support
• Freshness, volume, and schema drift observability using zero-cost metadata signals — no full table scans required
• Criticality scoring using the same weighted framework as Prizm’s other connectors
• Auto-generated quality rules from existing Oracle constraints (NOT NULL, PK/UK, FK, CHECK)
• Table, view, procedure, and package lineage via foreign keys and dependency metadata — no SQL parsing required for the standard case
• Usage analytics from V$SQL/V$SQLAREA, extended by AWR where licensed
• Catalog integrations with Alation, Atlan, Azure Purview, and Collibra
• Exception record extraction to Apache Iceberg for downstream remediation and audit workflows

See the full Oracle connector documentation for setup and configuration details. [Screenshot: Oracle connector source detail form]

---

​

New Sigma Computing Connector

Prizm now connects to Sigma Computing, cataloguing Workbooks, Pages, Data Elements, and Datasets, with end-to-end lineage from warehouse tables through Sigma to end consumers. Key Capabilities

• Authentication via Client ID + Client Secret (recommended) or OAuth 2.0 through GCP IAP
• Full catalog extraction of Workspaces, Workbooks, Pages, Data Elements, and Datasets, with calculated fields distinguished from base fields and tracked for formula changes
• End-to-end lineage via the Sigma Logs API, Dataset warehouse-source metadata, and direct column references
• Criticality scoring (60% dependency, 40% usage) matching the framework used for Power BI and Tableau
• Freshness and refresh-health monitoring, with alerts for stale workbooks, failed refreshes, upstream schema changes, and unused workbooks
• DQ score propagation from upstream warehouse tables through Sigma Datasets to Workbooks
• Automatic Prizm issue creation for Critical/High severity Sigma events, with Jira/ADO cascade support

Sigma’s Chrome Extension, dedicated analytics views, and external catalog integrations (Alation, Atlan, Purview) are on the roadmap but not part of this release — usage and reliability metrics also depend on Premier-tier Sigma audit log access. See the full Sigma connector documentation for setup and configuration details.

---

​

Power BI Connector — Semantic Model Support

Prizm’s Power BI connector now catalogs Semantic Models (formerly “datasets”) as first-class assets, extending lineage, criticality, and metrics one layer upstream of Reports and Dashboards. Key Capabilities

• Full semantic model schema extraction via the XMLA endpoint — tables, columns, measures, calculated columns, relationships, hierarchies, calculation groups, and partitions
• Upstream lineage stitching from semantic model tables to cataloged Snowflake/Databricks assets via M-expression parsing, including resolution of named/shared source expressions
• Downstream lineage extended through to Reports and Dashboards, with dataflow hops included when enabled
• Criticality scoring for semantic models, weighted by their highest-scoring report, report count, and refresh reliability
• New Schema Drift Count metric, tracking upstream structural changes propagated into the semantic model
• Monitoring corrected to a 7-job pipeline (Technical, Operational, Performance, Profile/Structural Recommendation, Profile, Structural), replacing the earlier 3-job model that predated the semantic-model architecture

See the Power BI connector documentation for full details.

---

​

ADLS Connector — Unity Catalog Iceberg Materialization & Delta Lake Support

The ADLS connector now supports running quality jobs on your own Databricks Spark cluster instead of Prizm’s managed Spark engine, and — for clients on that path — can materialize selected ADLS assets as Unity Catalog–managed, Iceberg-compatible tables (via Delta UniForm) that any Iceberg REST–compatible engine can query directly, with no export step or second data copy. Key Capabilities

• Unity Catalog Iceberg materialization: ADLS files processed on a customer-owned Databricks cluster are written as persistent, governed Delta tables with Iceberg compatibility — queryable by Trino, external Spark, Snowflake, or DuckDB via the metastore’s Iceberg REST endpoint
• Two supported authentication paths for Azure — Managed Identity (recommended, no secret to rotate) or Service Principal — each scoped across a read-only source container and a separate read-write managed-storage container
• Native Delta Lake table discovery: folders containing a _delta_log now catalog as a single table asset instead of one asset per Parquet part file, fixing prior format-read errors and false corruption signals
• Exception JSON staging: for ADLS + Databricks Unity Catalog clients, exception and metadata instruction JSON now stages to the client’s own write external location rather than uploading to Prizm SaaS S3 or via scheduled Azure Blob upload
• Five compute metric types supported on customer Databricks compute: Conditional, Query, Lookup, Comparison, and Behavioral
• A consolidated grant script and pre-go-live validation checklist for Azure/Databricks admins standing up the integration

See the full ADLS connector documentation for setup and configuration details. [Screenshot: Unity Catalog external location and catalog setup in Databricks]

---

​

Exception Management Enhancements

​

Exception by Metric — Grouped View

Exception records can now be viewed grouped by the metric that generated them, alongside the existing flat list view. Use the view switcher on the Exception page to toggle between All Exceptions and Exception by Metric. Each metric group rolls up its records into a single row showing the metric name, the owners across its records (as stacked avatars), and a summary such as “2 Total · 0% Breached.” Expanding a group reveals its individual records with a dedicated column set: Exception, Failing Column, Owner (inline-assignable), Status, Age, SLA, and Resolution Note. Two sub-tabs — All and Conditional — let you isolate records tied to conditional threshold logic, and an Export action downloads the current grouped view. [Screenshot: Exception by Metric grouped view]

​

Field-Level Failure Breakdown

Opening an individual exception record now shows a full row-by-row breakdown of the failing data, not just a summary: Field, Current Value, Previous Value, Related Criteria, Action (Valid / Out of Range / etc.), and Comment — with the specific field(s) that caused the failure visually highlighted. An Action Items panel tracks remediation steps against the record, with AI-assisted suggestions available.

​

Exception Audit Log

Every exception record now has a dedicated Audit tab — an immutable, timestamped history of every status change, owner reassignment, and SLA transition (e.g. “SLA status updated from breached to stopped”), each entry recording the acting user and, where available, client IP and device.

​

Real-Time Exception Analytics Dashboard — AI Summary & Trend Enhancements

The Exception Analytics dashboard (Analytics → Exception) now opens with an AI-generated summary calling out the most important signal in the current view — for example, a growing backlog or a slow resolution rate — and points to the relevant widget. The exception trend chart now tracks three lines instead of two: Open, Resolved, and Auto Resolve, making it possible to see how much of your resolution rate is manual versus automatic. See the Exception Records and Exception Dashboard documentation for the full reference. [Screenshot: Exception Analytics dashboard with AI Summary panel]

---

​

Actions — Workflow Automation

Prizm’s Actions module lets you build configurable trigger → condition → action rules that automate routine work across the platform — escalating breached SLAs, assigning owners, sending notifications, and more — without someone manually watching for each event. Key Capabilities

• Visual, canvas-based rule builder with a linear top-to-bottom flow, plus a JSON view of the same rule for direct editing
• Generate AI: create a starting rule from a natural-language prompt
• Validate: check a rule for configuration errors before saving
• Trigger nodes scoped to a target entity (or all entities), with one or more conditions that must be met to proceed
• ELSE IF condition branching, so a single rule can route to different outcomes rather than a single linear path
• Action nodes including Assign Owner (with an assignment mode and per-role assignees across Business, Technical, and Steward) and Send Notification, runnable in sequence
• Per-rule Execution history (execution ID, status, duration, actor, action name, nodes touched) and a separate Audit log of every change to the rule’s own configuration
• Rules move through a Pending → Verified lifecycle and can be enabled/disabled without deletion
• Scoped today to Exception, Alert, Asset, Domain, Job, Schedule, SLA, and Term rule types

See the Actions documentation for the full rule-builder reference. [Screenshot: Actions rule builder canvas]

---

​

Issue Management Enhancements

​

AI Recommended Actions — Now Trackable

Recommended Actions on an issue are no longer static text. Each AI-generated next step now has its own checkbox so it can be marked complete individually, a refresh control regenerates the list on demand, and a + control lets you add a step manually.

​

Resolution & Verification Workflow

Issue Overview now includes Reason Code and Resolution Note fields alongside Status, Priority, and Assigned To — giving resolution a structured record rather than just a status flip. Issues also carry an Expected Resolution date derived from their bound SLA policy, and can link out to external tickets (Jira, ServiceNow) shown directly on the Overview under Integrations.

​

Issue Lifecycle — Blocked and Reopened

The Issue list now shows five live status cards — New, In Progress, Blocked, Resolved, and Reopened — up from three. A Reopened issue indicates its underlying alert re-fired after the issue was previously resolved.

​

New Lineage Tab

Issue detail now has a dedicated Lineage tab, separate from Overview, showing affected assets and reports plus an interactive lineage graph with per-node health, row/column counts, and alert/issue activity badges. See the Issues documentation for the full detail-page reference.

---

​

SLA

Alerts and Issues now carry a live SLA Status (Breached, Stopped) and, on Issues, an Expected Resolution target date driven by a shared SLA policy engine — plus synchronization to linked Jira and ServiceNow tickets. Covers the ServiceNow SLA integration for this release; see the SLA documentation for details.

---

​

Client Enhancements

​

Bulk Threshold Configuration for Metrics

The Metrics page’s More menu now includes Import (bulk-upload threshold values across many metrics via file) and Bulk Edit (apply a change to a selected set of metrics at once), in addition to existing Select Columns and Download options.

​

Custom Threshold Condition Fixes

Custom Threshold conditions now consistently expose the full set of comparison operators (equals, greater than, less than, between, etc.) regardless of whether the underlying field is numeric or text-typed — fixing a case where operators silently changed between the first and second time a condition was configured.

​

Clone Metric

Any metric can now be cloned from its row-level actions, making it easy to stand up a variant of an existing metric (for example, the same check scoped to a different lookback window) without rebuilding it from scratch. See the Metrics documentation for the full reference on bulk configuration and cloning.

​

User Display When Adding Group Members

Adding a user to a group via the + control in Access → Groups now shows the user’s full email address immediately, instead of showing only initials until the page was refreshed.

​

Analytics — Quality Score Label

The Analytics → Quality heading now reads “Latest Average Score” instead of “Current Score.”

---

​

Security

​

Vault Integration Across Configuration Pages

Vault-managed credentials can now be selected on additional configuration surfaces — Catalog Integration and Storage Configuration pages — alongside the source-connector pages that already supported it. Only the vault reference is persisted; secret values are never stored in the platform database or shown in plain text after save.

---

​

License Management

Continued work on the internal Prizm Control Hub, covering license creation and lifecycle management. This tooling is for internal DQLabs use only and is not part of the customer-facing Prizm application, so implementation details are intentionally not documented here.

---