Microsoft Fabric sets default Item Recovery to protect enterprise data assets from accidental deletion
Source: Microsoft
Starting August 16, 2026, Microsoft will enable Item Recovery by default across all Microsoft Fabric tenants that have not explicitly defined a custom retention policy. Supported items will automatically gain a three-day soft-delete protection window without requiring manual administrator setup. Existing custom configurations and explicitly disabled states remain completely untouched. This update shifts the platform toward safer operational defaults, cutting down support overhead and business disruptions caused by human or programmatic mistakes.
Automated safety net reduces platform operational risks
Accidental deletions are common in fast-moving enterprise data environments, especially when automation scripts, shared workspaces, and frequent CI/CD deployments are active. Previously, an unconfigured tenant offered no safety net, turning a simple misstep into a critical incident requiring complex recovery operations. Under the new default, deleted supported items move to a soft-deleted state within the Workspace Recycle Bin. Authorized teams can quickly self-serve and restore these items using native product workflows before hard deletion occurs after the retention window closes.
Enterprise governance teams retain complete control over retention and costs
This change establishes a baseline security standard while preserving full administrative sovereignty. Tenant admins can adjust the retention window between 3 and 90 days to align with corporate data protection policies, or opt out entirely via the Fabric admin portal. From a cost perspective, soft-deleted items incur standard OneLake storage charges during their retention period without introducing new licensing meters or hidden fees. Platform teams should review their settings and update internal incident runbooks prior to rollout.
Business value and strategic impact for enterprise organisations
For mid-market and enterprise data leaders, this update converts potential high-severity outage incidents into minor, routine support tasks. Enabling item recovery by default enforces proactive data resiliency across every workspace without adding administrative burden or requiring explicit setup for new tenants. By establishing a guaranteed recovery baseline, organisations strengthen platform governance, protect business continuity during rapid development cycles, and eliminate preventable downtime caused by accidental data destruction.
Learn more on Microsoft Fabric Updates Blog
Fluent 2 in Power BI and how to quickly restore traditional matrix visual styles
Source: Marco Russo/SQLBI
Microsoft’s new Fluent 2 design language introduces a minimalist, borderless layout as the default setting for Power BI Table and Matrix visuals. While this modernization delivers a clean aesthetic, it strips away functional features like alternating row shading and clear cell borders, making dense financial reports harder to read. Organizations updating report templates can easily fix individual matrices via the Format pane, but scaling this adjustment across enterprise datasets presents an unexpected scripting hurdle.
Standardizing matrix legibility across enterprise reports
The shift to Fluent 2 sets the default Style preset of newly created Matrix visuals to “None”. For reporting teams managing multi-page dashboards, alternating row colors (zebra striping) and visible gridlines provide essential visual anchors for tracking complex tabular data across wide screens.
Manually updating every visual through Power BI Desktop requires navigating to Format > Visual > Style presets and switching back to “Default”, a process that introduces significant manual overhead across large environments.
Automated formatting fixes require precise TMDL property handling
Automating this style restoration via script is surprisingly nuanced. Programmatic report updates using modern LLM code assistants often hit friction because Fluent 2 alters how visual container properties are declared in Tabular Model Definition Language (TMDL) and report layout JSON structures. Developers attempting to push bulk visual fixes must explicitly target the underlying visual style properties rather than relying on global theme overrides, making precise automated script generation essential for rapid remediation.
Restoring standard visual formatting ensures that financial analysts, operational managers, and executive stakeholders maintain optimal readability without sacrificing report accessibility standards. For data platform teams and Power BI developers, establishing clear visual theme policies early prevents redundant manual refactoring, keeping engineering time focused on data pipeline scalability and governance rather than UI polish.
Learn more on SQLBI
Microsoft Fabric simplifies SQL Server migration with new wizard-driven assistant
Source: Microsoft
Microsoft has introduced a Fabric-native Migration Assistant in Public Preview, removing the primary technical and operational barriers to moving legacy SQL Server workloads into the cloud. By pairing automated DACPAC schema conversion with integrated Data Factory pipelines, mid-market and enterprise organisations can now migrate database workloads directly into Fabric without complex refactoring or custom ETL construction.
Automated schema conversion reduces migration risk
Migrating legacy database schemas often requires months of manual scripting and compatibility testing. The new wizard-driven assistant automates this workflow by accepting standard DACPAC files, evaluating object compatibility, and highlighting unsupported features before any deployment begins. Built-in Copilot integration suggests interactive fixes for incompatible code, while pre- and post-deployment script hooks handle constraint management cleanly during data loads. This drastically shortens assessment phases and prevents common schema deployment failures.
Direct OneLake integration accelerates enterprise AI and analytics
The real architectural value lies in what happens after deployment. SQL database in Fabric leverages serverless compute with automatic scaling, high availability, and native vector search capabilities out of the box. Crucially, operational data automatically replicates to OneLake in near real-time without requiring duplicate storage or manual ingestion pipelines. Enterprise data teams can instantly expose live operational data to Power BI reports, Spark notebooks, and Azure OpenAI models while retaining existing T-SQL, SSMS, and VS Code workflows.
For data leaders managing aging SQL Server infrastructure or high-cost cloud databases, this preview offering fundamentally changes the business case for platform modernization. By removing the need for manual ETL re-architecture and reducing manual migration labor, organisations can consolidate operational data and enterprise analytics under a single, governed Fabric environment with significantly reduced operational overhead.
Learn more on Microsoft Fabric Updates Blog
Associated identities in Microsoft Fabric eliminate user dependencies for enterprise workloads
Source: Microsoft
Microsoft Fabric has introduced Public Preview support for assigning associated identities – including service principals and managed identities – to Lakehouses and Eventstreams. By decoupling item-level execution and connection access from individual user accounts, organisations can now run critical data pipelines without the risk of downtime caused by employee offboarding or credential expiration.
Automated identity management replaces fragile user ownership
Historically, core Fabric features like SQL endpoint delegated mode relied on the creator’s or owner’s personal identity. When a developer left the organisation or their password expired, connected workflows frequently failed, requiring manual ownership takeovers via the UI.
Through new REST API endpoints, platform teams can now programmatically assign, update, and inspect default identities for Lakehouse and Eventstream items, seamlessly integrating identity lifecycle management into existing enterprise DevOps pipelines.
Improved platform resilience and lower operational risk for enterprise teams
This shift directly benefits Data Platform Managers, Data Architects, and Lead Data Engineers focused on operational continuity. By removing reliance on named user accounts, enterprise data teams can enforce strict zero-trust principles and avoid emergency remediation drills when staffing changes occur.
While currently limited to Lakehouses and Eventstreams (excluding streams sourced from Azure or Fabric Events), this capability represents a vital step toward production-grade stability across the entire Fabric ecosystem. Enterprise IT leaders should audit current Lakehouse and Eventstream assets now and migrate high-priority items to managed identities or service principals.
Learn more on Microsoft Fabric Updates Blog
Microsoft Fabric improves Spark governance with capacity-level concurrency monitoring
Source: Microsoft
Microsoft Fabric now offers complete visibility into Spark job queuing and compute consumption across entire capacities. By expanding workspace-level monitoring into a unified, capacity-wide view, enterprise platform administrators can pinpoint the root cause of job throttling, eliminate cross-team resource contention, and optimize existing compute investments without increasing licensing costs.
Resolving cross-workspace resource contention
When Spark pipelines stall or return HTTP 429 throttling errors, data engineering teams often assume their local workspace configurations are at fault. In shared enterprise environments, however, delays are frequently caused by neighboring workloads peaking simultaneously on the same capacity. The new capacity-level view aggregates real-time compute unit (CU) consumption across all attached workspaces, allowing platform teams to distinguish between localized workspace pool limits, overall capacity saturation, and noisy-neighbor behavior in a single dashboard.
Identifying historical trends to optimize workload scheduling
To support proactive capacity management, administrators can track CU usage over 24-hour, 7-day, and 30-day windows. Short-term views simplify incident root-cause analysis during unexpected pipeline failures, while longer historical windows surface recurring patterns such as overlapping nightly ETL runs or month-end processing bottlenecks. Armed with these metrics, data platform managers can reschedule resource-heavy pipelines to off-peak hours, adjust workspace pool caps, or selectively enable auto-scale only when sustained demand requires it.
Maximizing compute efficiency and operational predictability
For enterprise data leaders balancing tight platform budgets with demanding performance SLAs, this capability shifts capacity management from reactive troubleshooting to data-driven governance. Gaining clear visibility into actual CU utilization prevents unnecessary capacity upgrades, ensures critical business pipelines run predictably, and maximizes the return on existing Microsoft Fabric investments.
Learn more on Microsoft Fabric Updates Blog
Power BI semantic models meet Copilot in Excel for direct business analysis
Source: Chris Webb’s BI Blog
Organizations can now analyze Power BI semantic models directly inside Excel using Microsoft Copilot. By combining centralized enterprise data models with familiar spreadsheet formulas, business users gain direct access to governed data without relying on manual data exports or custom report development. This integration bridges the gap between enterprise BI and everyday analytics, allowing teams to query models using natural language while executing calculations natively in Excel.
Shifting from DAX to native Excel calculations
When prompted to run complex tasks like an ABC analysis against a Power BI semantic model, Copilot extracts the required underlying dataset and translates the analytical logic into native Excel formulas rather than relying strictly on DAX execution. While this delivers immediate transparency for business users who understand spreadsheet logic better than DAX, it currently generates a static data snapshot. For platform architects and BI managers, the immediate value lies in lowered barriers to self-service analytics, though automated query connection refreshes remain a critical future requirement for enterprise workflows.
Accelerating corporate reporting and self-service adoption
Beyond raw analysis, Copilot supports enterprise formatting standards through skills like brandkit, allowing users to apply corporate themes and logos instantly with simple prompts. By meeting users directly in Excel, this capability drastically reduces reliance on unmanaged “Export to Excel” workarounds and ad-hoc report requests to central data teams. Data leads benefit from higher engagement with governed semantic models without managing fragmented local spreadsheet versions.
For enterprise data leaders, this update represents a pragmatic step toward scaling self-service BI without forcing business users into specialized reporting tools. By making central Power BI semantic models seamlessly accessible and understandable within Excel, organizations can maximize their existing data infrastructure investments, reduce reporting bottlenecks, and accelerate executive decision-making.
Learn more on Chris Webb’s BI Blog
