Azure Container Services Docs – Weekly Update (2026-08-23 to 2026-08-30)
The most meaningful Azure Kubernetes Service, AKS Application Network, Azure Arc-enabled Kubernetes, AKS Arc, Container Registry, Application Gateway for Containers, and Fleet Manager documentation changes from the last 7 days. AKS everywhere! Summaries are AI-filtered to skip trivial edits.
📊 Updates this week: ACR (4), AGC (6), AKS (22), Arc K8s (3), Fleet (1)
|
||
|
General ♻️ Rework Modified: 2026-08-28 17:03 |
||
|
Summary This update includes version bundle updates with new versions added (1.33.12, 1.34.8, 1.35.4, 1.35.5, 1.36.0, 1.36.1) which are significant for users needing the latest features and fixes in Azure Kubernetes Service (AKS). |
||
|
|
||
|
Networking ✨ Update Modified: 2026-08-25 15:13 |
||
|
Summary The documentation now clarifies that when creating an identity via the Add-on, the identity name will be `applicationloadbalancer-<cluster-name>`, providing users with specific naming conventions to expect. Additionally, it notes that the _AppGW for Containers Configuration Manager_ role is already assigned, which may streamline user setup processes. |
||
|
|
||
|
Compute ♻️ Rework Modified: 2026-08-27 06:03 |
||
|
Summary The documentation now clarifies that all VM sizes in a node pool must support the same ephemeral OS disk placement when using ephemeral OS disks. If a scale profile mixes VM sizes with different placements, the create or update operation will fail with an `InvalidParameter` error, specifically indicating that mixed placements are not supported. Users are advised to either select compatible VM sizes, separate them into different node pools, or opt for managed OS disks to avoid this issue. |
||
|
|
||
|
Networking ♻️ Rework Modified: 2026-08-27 18:43 |
||
|
Summary Updated the version numbers for the ALB Controller and Gateway API, reflecting the latest security patches and fixes. This ensures users have the most current information for deployment. |
||
|
|
||
|
General ♻️ Rework Modified: 2026-08-28 22:08 |
||
|
Summary The documentation for the `acr purge` command has been updated to include the new `–include-locked` parameter. This allows users with sufficient permissions to automatically unlock locked artifacts by resetting their `delete-enabled` and `write-enabled` statuses to `true` before deletion. For further details, users can refer to the section on automatically purging images from an Azure container registry. |
||
|
|
||
|
Security ♻️ Rework Modified: 2026-08-27 17:12 |
||
|
Summary The documentation now clarifies the steps to remove the Custom CA Trust property from node pools in Azure Kubernetes Service (AKS). Users are instructed to use a generic resource update instead of the previously available `–disable-custom-ca-trust` option, which is no longer included in current Azure CLI releases. Additionally, users can verify the successful update by querying the node pool resource to confirm that `enableCustomCATrust` is set to `false`. |
||
|
|
||
|
Storage ♻️ Rework Modified: 2026-08-26 06:02 |
||
|
Summary The documentation now includes a new section titled "Create a complete dynamically provisioned volume," which provides a detailed example for creating a general-purpose Azure Files storage class, a PersistentVolumeClaim (PVC), and a pod that mounts the file share using the `Standard_LRS` SKU and SMB protocol. Users can follow the step-by-step instructions to apply the manifest, confirm the PVC and pod states, and verify read/write access to the file share, enhancing their ability to utilize Azure Files in Kubernetes environments. |
||
|
|
||
|
Operations ♻️ Rework Modified: 2026-08-26 02:06 |
||
|
Summary Revised troubleshooting guide for errors encountered during AKS cluster upgrades and scaling operations. It includes updated descriptions and steps for identifying failed operations, improving clarity and usability for users facing these issues. |
||
|
|
||
|
Cost ♻️ Rework Modified: 2026-08-27 06:03 |
||
|
Summary The documentation now includes detailed guidance on running long-running batch jobs on Azure Kubernetes Service (AKS), emphasizing the importance of persisting progress outside the pod and handling termination signals. Users can implement checkpointing, retries, and termination strategies to enhance job reliability, with specific examples provided for configuring Kubernetes Jobs. Additionally, best practices for monitoring job progress and planning for node maintenance are outlined, ensuring users can effectively manage and recover from job interruptions. |
||
|
|
||
|
Operations ♻️ Rework Modified: 2026-08-26 06:02 |
||
|
Summary Updated the commands for listing auto-upgrade profiles and clarified the scope of Fleet auto-upgrade profiles. The changes include new commands and explanations for determining which profiles can update AKS clusters. |
||
|
|
||
|
General ♻️ Rework Modified: 2026-08-24 22:12 |
||
|
Summary Added details on Artifact Streaming, which allows streaming container images from Azure Container Registry (ACR) to AKS, reducing deployment times. It includes prerequisites and configuration steps for enabling this feature in Node Auto Provisioning clusters. |
||
|
|
||
|
General ♻️ Rework Modified: 2026-08-27 06:02 |
||
|
Summary The documentation for Azure Container Registry now clarifies that when importing images from another Azure Container Registry (ACR), users can only import from the source ACR’s home region and not from specific geo-replicas using regional endpoints. Additionally, it emphasizes that the source ACR’s global endpoint must be specified for imports, and all imports will write content to the downstream ACR’s home region. |
||
|
|
||
|
Networking ♻️ Rework Modified: 2026-08-24 22:12 |
||
|
Summary The documentation has been updated to reflect the latest Cilium versions compatible with Azure Kubernetes Service (AKS). Users can now see that Cilium versions 1.16.19, 1.17.18, 1.18.12, and 1.19.6 are available for various LTS releases, ensuring they have access to the most recent features and security updates. This change enhances the clarity of version compatibility for users deploying Cilium on AKS. |
||
|
|
||
|
Compute ♻️ Rework Modified: 2026-08-26 06:02 |
||
|
Summary Added examples for dedicating node pools to multiple applications, including commands for creating node pools with unique taints and labels. This enhances the documentation for managing workloads in AKS. |
||
|
|
||
|
General ♻️ Rework Modified: 2026-08-25 06:37 |
||
|
Summary Corrected links to the Kata Containers documentation, ensuring users have access to the correct resources for technical information. |
||
|
|
||
|
Security ♻️ Rework Modified: 2026-08-26 22:18 |
||
|
Summary Expanded guidance on network security practices in AKS, detailing recommended approaches for different service types and their security considerations. This update aims to improve the security posture of AKS deployments. |
||
|
|
||
|
Networking ♻️ Rework Modified: 2026-08-27 18:43 |
||
|
Summary The documentation now includes a note that the Application Gateway for Containers AKS add-on is currently in preview, along with a link to the Supplemental Terms of Use for Azure Previews. Additionally, users are instructed to use the Azure CLI to register the add-on feature, which is essential for provisioning the Application Gateway for Containers resources in Azure. The verification steps for the ALB Controller installation have been clarified, ensuring users understand the checks required after enabling the add-on. |
||
|
|
||
|
Security ♻️ Rework Modified: 2026-08-26 11:02 |
||
|
Summary Added new release note entries detailing patch releases that update dependencies with known vulnerabilities, enhancing security awareness for users. |
||
|
|
||
|
General ♻️ Rework Modified: 2026-08-27 22:09 |
||
|
Summary The documentation for deploying CanIPull to an AKS cluster has been updated to clarify that the tool now checks DNS resolution and the initial Azure Container Registry (ACR) token exchange from a selected node. Users are informed that CanIPull does not perform an end-to-end image pull test, and a successful result does not confirm access to specific repositories or images. Additionally, the deployment instructions have been refined to emphasize the use of the complete manifest and the specific parameters required for deployment. |
||
|
|
||
|
Operations ♻️ Rework Modified: 2026-08-29 06:06 |
||
|
Summary The documentation for the `–keep` parameter has been updated to clarify that it specifies the latest number of to-be-deleted tags or manifests per repository. Additionally, the new `–include-locked` parameter is introduced, which allows purging of locked tags and manifests, overriding their deletion protection. A caution note emphasizes that deleted data is unrecoverable, and users are advised to run with `–dry-run` first. |
||
|
|
||
|
Compute 🆕 New Modified: 2026-08-26 17:11 |
||
|
Summary Learn how to analyze GPU profiling flame graphs to identify functions that consume the most GPU memory in workloads on Azure Kubernetes Service. |
||
|
|
||
|
General ♻️ Rework Modified: 2026-08-26 22:18 |
||
|
Summary The documentation now includes enhanced guidance on migrating workloads to Azure Kubernetes Service (AKS) and releasing application updates with minimal or zero downtime. Users can now perform recurring application releases within an existing AKS cluster without needing a second cluster or DNS cutover, utilizing a Kubernetes `Deployment` with a `RollingUpdate` strategy. Additionally, the quick runbook has been updated to emphasize the use of relative DNS weights and precise request percentages for traffic shifting, improving the clarity of the migration process. |
||
|
|
||
|
General ♻️ Rework Modified: 2026-08-26 19:15 |
||
|
Summary The documentation now highlights significant changes in the `microsoft.flux` version 1.25.0, including the transition of the `helm-controller` from Helm v3 to Helm v4, which alters default behaviors such as server-side apply and kstatus-based health checks. Users are advised to enable the `useHelm3Defaults` setting to restore previous Helm v3 behaviors. Additionally, deprecated APIs have been removed, and users must update their sources to avoid disruptions in functionality. |
||
|
|
||
|
Networking ♻️ Rework Modified: 2026-08-27 18:43 |
||
|
Summary Updated version numbers in the quickstart guide for deploying the application gateway for containers, ensuring users are directed to the latest versions. |
||
|
|
||
|
Security ♻️ Rework Modified: 2026-08-25 22:09 |
||
|
Summary The documentation now clarifies that specific roles grant permissions to manage and view repository-level policies for images and artifact streaming configurations. Users can now understand that certain roles explicitly allow them to manage repository-level policies and view configurations related to artifact streaming, enhancing their ability to control and monitor these aspects within the container registry. |
||
|
|
||
|
General ♻️ Rework Modified: 2026-08-24 22:14 |
||
|
Summary The documentation now includes important updates regarding the `helm-controller` in `microsoft.flux` v1.25.0, which has transitioned to using Helm v4, introducing changes in default behaviors such as server-side apply and kstatus-based health checks. Users can restore Helm v3 default behavior by using the `–config-settings helm-controller.useHelm3Defaults=true` command, allowing for compatibility with existing Helm releases while they validate against the new Helm v4 defaults. |
||
|
|
||
|
General ♻️ Rework Modified: 2026-08-24 22:12 |
||
|
Summary Users can now enable artifact streaming for clusters with Node Auto Provisioning (NAP) by utilizing the `spec.artifactStreaming.enabled` field in the AKSNodeClass CRD. By setting this field to `true`, users can activate artifact streaming for both new and existing NAP-managed nodes. Additionally, the section title for enabling artifact streaming on an existing node pool has been updated for clarity. |
||
|
|
||
|
Networking ♻️ Rework Modified: 2026-08-27 18:43 |
||
|
Summary The ALB Controller has been updated to version 1.11.4, which includes security patches and introduces the Gateway API Inference Extension Version 1.3.1. Users should note that the minimum Kubernetes version remains at v1.27, and the previous version 1.11.3 continues to include security patches, WAF fixes, and an Istio service mesh integration cert-rotation fix. |
||
|
|
||
|
General ♻️ Rework Modified: 2026-08-28 21:09 |
||
|
Summary The documentation for Azure Kubernetes Service (AKS) has been updated to clarify the division of responsibilities between Microsoft and users regarding support policies. A new responsibility matrix has been added, detailing what Microsoft manages versus what users are responsible for, including aspects like node configuration, workloads, and networking. Additionally, the article now specifies that Microsoft Support has limited access to agent nodes and emphasizes the importance of using Kubernetes-native mechanisms for changes to avoid rendering the cluster unsupportable. |
||
|
|
||
|
Networking ♻️ Rework Modified: 2026-08-25 13:05 |
||
|
Summary Updated documentation for configuring container network logs, including changes to namespace specifications for Kubernetes pods. |
||
|
|
||
|
Networking ♻️ Rework Modified: 2026-08-27 18:43 |
||
|
Summary The documentation now clarifies that if an association is deleted and recreated, it will be treated as a new association subject to the current NSG enforcement model. Users are advised to configure the necessary NSG allow rules, including the _AzureLoadBalancer_ tag, before recreating the association to prevent connectivity issues. Additionally, the section on routing patterns for the association subnet has been updated to emphasize the importance of disabling BGP route propagation and directing 0.0.0.0/0 traffic directly to the internet to avoid asymmetrical routing problems. |
||
|
|
||
|
Compute ♻️ Rework Modified: 2026-08-26 17:11 |
||
|
Summary The documentation for GPU profiling on Azure Kubernetes Service (AKS) has been updated to clarify the setup process. Users can now learn how to set up GPU memory profiling using Inspektor Gadget, Pyroscope, Azure Managed Grafana, and Azure Monitor managed service for Prometheus. Additionally, the article now emphasizes the importance of enabling GPU profiling before deploying workloads to capture memory allocation events effectively. |
||
|
|
||
|
Operations 🆕 New Modified: 2026-08-27 22:09 |
||
|
Summary Learn how to diagnose and resolve provisioning, power state, and node health issues in an Azure Kubernetes Service agent pool. |
||
|
|
||
|
Compute ♻️ Rework Modified: 2026-08-26 06:02 |
||
|
Summary The documentation now correctly reflects that Ubuntu 20.04 is the default OS version for Kubernetes versions 1.24 to 1.34, while Ubuntu 24.04 is the default for Kubernetes versions 1.35 to 1.38. This update ensures users have accurate information regarding the supported OS versions for their AKS deployments. |
||
|
|
||
|
General ♻️ Rework Modified: 2026-08-24 06:02 |
||
|
Summary The documentation for custom private DNS zones in AKS has been updated to clarify that AKS will create a virtual network link for the spoke VNet hosting the cluster, even if the DNS zone is already linked to a hub VNet. Users must ensure that the cluster’s managed identity has **Network Contributor** permissions on the spoke VNet before setting `publicDNS` to `false` or using `–disable-public-fqdn`. This change emphasizes the necessary permissions for proper configuration. |
||
|
|
||
|
General ♻️ Rework Modified: 2026-08-24 22:12 |
||
|
Summary The documentation for supported Kubernetes versions has been updated to reflect new versions of various components, including cilium-agent and cilium-operator, which have been upgraded to v1.19.6-260811. Additionally, the metrics-server has been updated to v0.8.0-13, and several KEDA components have been upgraded to versions 2.19.0-6 for the main KEDA service and related admission webhooks. Users can now leverage these updated components for improved functionality and performance in their Kubernetes environments. |
||
|
Full tracker with filters: Azure Container Services Docs Tracker
