Azure Kubernetes Service (AKS)

Have feedback for Azure Kubernetes Service (AKS)? Submit your idea or vote up an idea submitted by others. All of the feedback you share here will be monitored and reviewed by the AKS team. Remember that this site is only for feature suggestions and ideas. If you have technical questions or issues, please submit them to our GitHub issues page or on StackOverflow.

  • Hot ideas
  • Top ideas
  • New ideas
  • My feedback
  1. AKS Upgrade uncordon failed because of casing issue

    During the AKS Upgrade (1.13.5 -> 1.14.6) I got this error message :

    Error to uncordon the node aks-nodepool1-00000000-vmss00000B, nodes "aks-nodepool1-00000000-vmss00000B" not found

    This node does exist but with a lower case 'B'.

    Everytime we have a node ending with letters, this error happens.

    This is causing our nodes to be stuck in "SchedulingDisabled" and we have to manually uncordon them.

    Look at the attached file.

    1 vote
    Sign in
    (thinking…)
    Sign in with: Microsoft
    Signed in as (Sign out)

    We’ll send you updates on this idea

    1 comment  ·  CLI  ·  Flag idea as inappropriate…  ·  Admin →
  2. Allow selecting disk type when creating AKS cluster

    There is no way to select disk type (Standard or Premium) when creating an AKS cluster via Portal or Azure CLI, by default Node OS disk type is derived from the selected VM size.

    To be more specific, I want a node Standard_B2s with Standard disk. If I create a plain VM, I can select such configuration. If I create an AKS cluster with Standard_B2s node, I can stop the node and change the disk type to Standard, but I cannot select Standard_B2s + Standard disk while creating AKS cluster.

    https://github.com/Azure/AKS/issues/580

    53 votes
    Sign in
    (thinking…)
    Sign in with: Microsoft
    Signed in as (Sign out)

    We’ll send you updates on this idea

    4 comments  ·  CLI  ·  Flag idea as inappropriate…  ·  Admin →
  3. Fix incremental AKS ARM deployment when autoscaler enabled

    Currently, whenever you have the autoscaler enabled and you want to do a new deployment you need to dynamically provide the current sizings of the cluster (min / max / current no. of nodes) for the arm template to be valid. I think when autoscaler is already enabled on the target cluster, it should disregard these as a kind of initial settings.

    1 vote
    Sign in
    (thinking…)
    Sign in with: Microsoft
    Signed in as (Sign out)

    We’ll send you updates on this idea

    0 comments  ·  CLI  ·  Flag idea as inappropriate…  ·  Admin →
  4. AKS Upgrade drain timeout caused by local storage

    When trying to upgrade (1.13.5 -> 1.14.6) with local storage, I'm getting a drain timeout after 20mins (per node).

    This is causing my upgrade to take about 4 hours.

    If would be a nice improvement if the drain command could contain the --delete-local-data flag or give us the possibility to specify this flag for the az aks upgrade.

    1 vote
    Sign in
    (thinking…)
    Sign in with: Microsoft
    Signed in as (Sign out)

    We’ll send you updates on this idea

    0 comments  ·  CLI  ·  Flag idea as inappropriate…  ·  Admin →
  5. AKS Upgrade drain timeout caused by DaemonSet

    When trying to upgrade (1.13.5 -> 1.14.6) with node that has DaemonSet, I'm getting a drain timeout after 20mins (per node).

    This is causing my upgrade to take about 4 hours.

    If would be a nice improvement if the drain command could contain the --ignore-daemonsets flag or give us the possibility to specify this flag for the az aks upgrade.

    1 vote
    Sign in
    (thinking…)
    Sign in with: Microsoft
    Signed in as (Sign out)

    We’ll send you updates on this idea

    0 comments  ·  CLI  ·  Flag idea as inappropriate…  ·  Admin →
  6. Recreate cluster nodes

    During the upgrade process it tears down all existing VM's and assigns new VHD/Tags etc.

    But this process seems to be only available for az aks upgrade, what i would like to be able to do is to refresh all the vms and not rely on manually deleting to purge them out of the cluster.

    Something like

    az aks upgrade -vm

    It skips the cluster upgrade without changing anything and just gets the latest VHD assignments and aks engine numbers.

    The background is, that i scaled the cluster to 5 Nodes and 4 were on aks 0.35 and the node0…

    2 votes
    Sign in
    (thinking…)
    Sign in with: Microsoft
    Signed in as (Sign out)

    We’ll send you updates on this idea

    1 comment  ·  CLI  ·  Flag idea as inappropriate…  ·  Admin →
  7. Add way to query which Feature-Flags are enabled

    From: https://github.com/Azure/AKS/issues/187

    There are workloads that are starting to show up that require certain Kube-api, controller-manager, Scheduler feature flags to be enabled to deploy a specific way. Examples:
    Istio with auto side-car injection - aggregated-API flag
    RBAC - RBAC Flag
    HPA based on customer metrics - aggregated API
    Auditing - audit policy flags

    Perhaps a cli that can list the different flags enabled on the control plane components that are running on a specific cluster:

    az aks get feature-flags -n <aks-resource-name> -g <resourceGroup Name>

    Agin this would just be to query and not change so in selecting possible workloads it…

    29 votes
    Sign in
    (thinking…)
    Sign in with: Microsoft
    Signed in as (Sign out)

    We’ll send you updates on this idea

    0 comments  ·  CLI  ·  Flag idea as inappropriate…  ·  Admin →
  8. AZ AKS Create command creates a new resource group

    The az aks create command allows you to specify a resource group to deploy to. However, this resource group only gets the kubernetes service, the command then creates a NEW RG that starts with MC_. This RG does not have the same permissions as the originally specified RG and therefore the user cannot create additional services in the new RG, such as ALB or PIP.

    8 votes
    Sign in
    (thinking…)
    Sign in with: Microsoft
    Signed in as (Sign out)

    We’ll send you updates on this idea

    0 comments  ·  CLI  ·  Flag idea as inappropriate…  ·  Admin →
  9. Support for cluster up and down commands

    Sometimes it helps to shutdown and restart the entire cluster with a single command like OpenShift does. It help purging some objects that are in failed state.

    9 votes
    Sign in
    (thinking…)
    Sign in with: Microsoft
    Signed in as (Sign out)

    We’ll send you updates on this idea

    0 comments  ·  CLI  ·  Flag idea as inappropriate…  ·  Admin →
  10. Ability to know new AKS version is released, via Az CLI or Powershell Modules

    Az CLI or Powershell modules to provide new AKS versions available.

    1 vote
    Sign in
    (thinking…)
    Sign in with: Microsoft
    Signed in as (Sign out)

    We’ll send you updates on this idea

    0 comments  ·  CLI  ·  Flag idea as inappropriate…  ·  Admin →
  11. Add Method for querying available regions from the Azure CLI

    It would be nice to be have the ability to list the available regions for AKS directly from the CLI. Something like:

    az aks regions list -o table

    2 votes
    Sign in
    (thinking…)
    Sign in with: Microsoft
    Signed in as (Sign out)

    We’ll send you updates on this idea

    0 comments  ·  CLI  ·  Flag idea as inappropriate…  ·  Admin →
  • Don't see your idea?

Feedback and Knowledge Base