DailyAzureUpdatesGenerator

October 01, 2025 - Azure Updates Summary Report (Details Mode)

Generated on: October 01, 2025 Target period: Within the last 24 hours Processing mode: Details Mode Number of updates: 4 items

Update List

Published: September 30, 2025 19:30:35 UTC Link: Public Preview: Soft Delete feature in Azure Compute Gallery

Update ID: 506886 Data source: Azure Updates API

Categories: In preview, Compute, Azure VM Image Builder, Virtual Machine Scale Sets, Virtual Machines, Gallery, Features, Services

Summary:

For more details, visit: https://azure.microsoft.com/updates?id=506886

Details:

The newly introduced Soft Delete feature in Azure Compute Gallery, currently in public preview, addresses a critical operational risk by enabling the recovery of accidentally deleted images within a configurable retention window of 7 days. This enhancement is designed to improve image management resilience and operational continuity for IT professionals managing large-scale VM image repositories.

Background and Purpose:
Azure Compute Gallery (formerly Shared Image Gallery) is a service that facilitates the management, sharing, and versioning of VM images across subscriptions and regions. Prior to this update, deletion of images was immediate and irreversible, posing a risk of data loss due to human error or automation faults. The Soft Delete feature was introduced to mitigate this risk by providing a safety net that allows administrators to recover deleted images, thereby reducing downtime and rework associated with image recreation.

Specific Features and Detailed Changes:

Technical Mechanisms and Implementation Methods:
Soft Delete is implemented as a state transition within the Azure Compute Gallery’s backend metadata store. Upon deletion, the image’s metadata is flagged as soft deleted, and the actual image blobs remain intact but inaccessible for provisioning. The system maintains a retention timer, after which a background process permanently deletes the image data. Recovery operations update the metadata state back to active, restoring normal access. This approach ensures minimal disruption to storage and indexing systems while providing a reversible deletion workflow.

Use Cases and Application Scenarios:

Important Considerations and Limitations:

Integration with Related Azure Services:
Soft Delete in Azure Compute Gallery complements Azure Backup and Azure Resource Manager’s (ARM) resource locks by providing a focused recovery mechanism for VM images. It integrates seamlessly with existing image versioning and replication capabilities, ensuring that soft delete states are respected across regional replications. Additionally, it aligns with Azure Policy and governance frameworks by enabling safer image lifecycle management within enterprise environments.

In summary, the Soft Delete feature in Azure Compute Gallery enhances image lifecycle management by introducing a recoverable deletion state with a 7-day retention window, thereby reducing operational risks associated with accidental deletions and improving overall service reliability for IT professionals managing VM images at scale.


2. Retirement: Support for Service Connector (preview) on Azure Container Apps will end on March 30th, 2026

Published: September 30, 2025 17:30:02 UTC Link: Retirement: Support for Service Connector (preview) on Azure Container Apps will end on March 30th, 2026

Update ID: 501528 Data source: Azure Updates API

Categories: Retirements

Summary:

Details:

The announced retirement of support for Service Connector (preview) on Azure Container Apps, effective March 30, 2026, signifies Microsoft’s planned discontinuation of this feature to streamline and evolve the integration capabilities within Azure Container Apps. Service Connector (preview) was introduced to simplify the process of connecting Azure Container Apps to various Azure services by automating network and security configurations, thereby reducing manual setup complexity and accelerating development workflows.

Background and Purpose:
Service Connector (preview) was designed as a low-code/no-code solution to enable seamless connectivity between containerized applications and Azure services such as Azure Database for PostgreSQL, Azure Cache for Redis, and Azure Event Hubs. The preview status indicated ongoing development and refinement based on user feedback. The retirement notice reflects Microsoft’s intent to phase out this preview feature, likely in favor of more robust, integrated, or evolved connectivity solutions within Azure Container Apps or other Azure networking services.

Specific Features and Detailed Changes:
The key feature of Service Connector (preview) was its ability to automatically establish secure connections by configuring managed identities, virtual network integrations, and firewall rules without requiring extensive manual intervention. With the retirement, these automated connection capabilities via Service Connector (preview) will cease to be supported. Users who have created service connections through the Azure Portal using this feature will need to plan migration strategies before the cutoff date to avoid disruption.

Technical Mechanisms and Implementation Methods:
Service Connector (preview) functioned by orchestrating underlying Azure resources and permissions: it leveraged Azure Managed Identities for authentication, configured virtual network service endpoints or private endpoints for secure connectivity, and adjusted firewall settings to permit traffic from Azure Container Apps to the target service. The implementation abstracted these complex configurations into a simplified UI-driven experience. Post-retirement, users will need to manually configure these components or use alternative automation mechanisms such as Azure CLI, ARM templates, or Terraform scripts to establish similar connectivity.

Use Cases and Application Scenarios:
Typical use cases included microservices architectures deployed on Azure Container Apps that required secure, reliable access to backend services like databases, caches, or messaging systems. For example, a containerized web API connecting to Azure Database for PostgreSQL or an event-driven application consuming Azure Event Hubs would benefit from Service Connector’s automated setup. With the retirement, these scenarios remain valid but will require more hands-on configuration.

Important Considerations and Limitations:

Integration with Related Azure Services:
Service Connector (preview) tightly integrated with Azure networking constructs such as virtual networks, private endpoints, and Azure Active Directory managed identities. It also interfaced with Azure services like Azure Database for PostgreSQL, Azure Cache for Redis, Azure Event Hubs, and others to streamline connectivity. Post-retirement, users should leverage Azure Private Link, virtual network integration for Azure Container Apps, and Azure RBAC to manually replicate these integrations. Additionally, Azure DevOps or GitHub Actions can be used to automate deployment and configuration of these network and identity components.

In summary, the retirement of Service Connector (preview) on Azure Container Apps by March 30, 2026, requires IT professionals to transition from this automated preview feature to manual or alternative automated methods for securely connecting containerized applications to Azure services, ensuring continuity and security in their cloud-native architectures.


3. Retirement: NVv4-series Azure Virtual Machines will be retired on September 30, 2026

Published: September 30, 2025 17:15:32 UTC Link: Retirement: NVv4-series Azure Virtual Machines will be retired on September 30, 2026

Update ID: 500578 Data source: Azure Updates API

Categories: Retirements

Summary:

Details:

The NVv4-series Azure Virtual Machines, including Standard_NV4as_v4, Standard_NV4ahs_v4, Standard_NV8as_v4, Standard_NV8ahs_v4, Standard_NV16as_v4, Standard_NV16ahs_v4, Standard_NV32as_v4, and Standard_NV32ahs_v4, are scheduled for retirement on September 30, 2026. This planned decommissioning reflects Microsoft’s ongoing efforts to optimize its VM portfolio by phasing out older generation instances in favor of newer, more efficient, and higher-performing options.

Background and Purpose
The NVv4-series VMs were introduced to provide GPU-accelerated compute capabilities tailored for graphics-intensive workloads, such as visualization, simulation, and remote workstation scenarios, leveraging AMD Radeon Instinct MI25 GPUs with AMD EPYC CPUs. Over time, advancements in GPU technology, VM architecture, and Azure infrastructure have led to the availability of newer VM series that offer improved performance, cost efficiency, and feature sets. The retirement aims to streamline the Azure VM offerings, reduce maintenance overhead, and encourage customers to migrate to more modern VM types that better meet current and future workload demands.

Specific Features and Detailed Changes
The NVv4-series VMs uniquely combined AMD EPYC CPUs with AMD Radeon GPUs, supporting GPU partitioning via AMD’s Multiuser GPU (MxGPU) technology, enabling multiple users to share a single GPU securely and efficiently. These VMs supported a range of sizes from 4 to 32 vCPUs, with corresponding GPU resources, designed for workloads such as CAD, 3D rendering, and video editing.

With the retirement, these specific VM SKUs will no longer be available for provisioning or scaling after the cutoff date. Existing deployments must be migrated or decommissioned before September 30, 2026, to avoid service disruption.

Technical Mechanisms and Implementation Methods
The NVv4-series leveraged AMD’s MxGPU technology implemented via SR-IOV (Single Root I/O Virtualization), allowing hardware-level GPU virtualization and secure multi-tenant GPU sharing. This differs from NVIDIA’s GRID technology used in other Azure GPU VMs. The retirement means that Azure’s backend infrastructure will cease support for these AMD GPU-based VMs, and the associated hardware will be phased out from Azure datacenters.

Migration paths typically involve moving workloads to newer GPU VM series, such as the NVv5-series or ND-series, which use NVIDIA GPUs and provide enhanced AI, HPC, and visualization capabilities. Migration may require application compatibility testing, driver updates, and possible reconfiguration of GPU-accelerated workloads.

Use Cases and Application Scenarios
NVv4-series VMs were primarily used for remote visualization, virtual desktop infrastructure (VDI), engineering simulations, and graphics rendering workloads that benefit from GPU acceleration. Organizations running CAD applications, 3D modeling, or media processing pipelines leveraged these VMs for cost-effective GPU sharing.

Post-retirement, customers should evaluate newer VM families that align with their workload requirements, considering factors such as GPU architecture, driver support, performance benchmarks, and cost. For example, NVv5-series VMs provide AMD MI250 GPUs optimized for HPC and AI, while ND-series VMs focus on NVIDIA A100 GPUs for deep learning and HPC.

Important Considerations and Limitations

Integration with Related Azure Services
NVv4-series VMs integrate with Azure services such as Azure Virtual Desktop, Azure Batch, and


Published: September 30, 2025 13:45:49 UTC Link: Retirement: Spark Native Connector for Semantic Link

Update ID: 502602 Data source: Azure Updates API

Categories: Analytics, Microsoft Fabric, Retirements

Summary:

Details:

The Azure update announces the official retirement of the Spark Native Connector for Semantic Link effective October 1, 2025. This decision follows an internal evaluation revealing minimal usage, disproportionate maintenance overhead, and identified security risks, prompting Microsoft to discontinue support and encourage migration to alternative solutions.

Background and Purpose:
The Spark Native Connector for Semantic Link was designed to facilitate seamless integration between Apache Spark workloads and Semantic Link services, enabling data scientists and engineers to query and analyze semantic graph data directly within Spark environments. However, over time, usage metrics indicated that the connector did not achieve widespread adoption, while ongoing maintenance and security patching demands increased operational costs. The retirement aims to streamline Azure’s service portfolio, reduce security exposure, and focus resources on more robust and widely used integration methods.

Specific Features and Changes:
The connector provided native Spark APIs and data source interfaces to connect Spark jobs with Semantic Link graph data stores, supporting read and write operations with optimized query translation. With retirement, the connector will no longer receive updates, bug fixes, or security patches after October 1, 2025, and will be formally deprecated before that date. Users will lose official support and should plan to transition workloads accordingly.

Technical Mechanisms and Implementation:
The connector functioned as a Spark DataSource V2 implementation, translating Spark SQL queries into Semantic Link graph queries and managing data serialization/deserialization between Spark’s RDD/DataFrame abstractions and the Semantic Link graph format. It leveraged Spark’s Catalyst optimizer extensions to push down predicates and optimize query execution. Its retirement means these integration points will cease to be maintained, and users must adopt alternative integration methods such as REST APIs, SDKs, or custom connectors.

Use Cases and Application Scenarios:
Primarily, the connector was used in scenarios where organizations performed large-scale semantic graph analytics within Spark clusters, such as knowledge graph enrichment, entity resolution, and relationship discovery in big data pipelines. It enabled data engineers to embed semantic querying directly into Spark ETL workflows. Post-retirement, these use cases will require re-architecting to use other integration approaches.

Important Considerations and Limitations:

Integration with Related Azure Services:
Users are encouraged to leverage Azure-native alternatives such as Azure Synapse Analytics, Azure Data Factory, or Azure Cosmos DB’s graph APIs for semantic data processing. Additionally, integration via RESTful APIs or Azure SDKs can replace the connector’s functionality, providing more secure and maintainable access to Semantic Link data. Combining these services with Azure Databricks or HDInsight Spark clusters can replicate or enhance existing workflows without relying on the deprecated connector.

In summary, the retirement of the Spark Native Connector for Semantic Link reflects Microsoft’s strategic shift to optimize service offerings by discontinuing low-usage, high-maintenance components with security concerns. IT professionals should proactively assess and migrate their Spark-based semantic graph workloads to supported integration methods leveraging Azure’s broader data ecosystem to ensure continuity, security, and performance beyond October 2025.


This report was automatically generated - 2025-10-01 03:02:25 UTC