-
Notifications
You must be signed in to change notification settings - Fork 318
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
[Retirement] Migrate from acs-mirror.azureedge.net to packages.aks.azure.com #4792
Comments
How will those changes be communicated? |
Thanks for your comment @ms1111. Information regarding any changes to the IP addresses will be available in AKS release notes. |
Since today we're getting alerts that requests from our k8s cluster to packages.aks.azure.com are getting blocked, but we do have the FQDN Tag "AzureKubernetesServices" whitelisted in our Azure Firewall. Is this update of the FQDN tag still pending or do we maybe have to act here? |
I think this service tag is referring to "after change", not "before change". If you check the IP result from two different domains, the deprecated one is not in the list. |
I've been seeing that since last week and am trying to persuade Azure support that there is an issue! |
@rahulrai-in the ServiceTag is currently not working, is this pending, or should it be completed already? I've raised a support ticket with Azure about this:
|
@toredash @rlaveycal @JoeyC-Dev @TomkhaLoL Thanks for the feedback. |
@rahulrai-in What is the timeline for the FQDN tag update ? |
Currently we are expecting to complete the roll out of the tag update in March'25. |
Due to the Edgio retirement, Azure Kubernetes Service (AKS) is updating its egress requirements. The FQDN
acs-mirror.azureedge.net:443
—currently used to download and install binaries for necessary dependencies (e.g., kubenet and Azure CNI)—will experience degraded performance starting March 31, 2025, and will be fully retired on May 31, 2025.Required action
If you use a Layer 7 (FQDN) firewall to restrict AKS egress traffic, action may be required:
1. Azure Firewall with AzureKubernetesService FQDN Tag
If you rely on Azure Firewall Application rules with the
AzureKubernetesService
FQDN tag, no action is needed. The update will be applied automatically.2. Other Firewalls or Egress Restrictions
If you restrict egress using another solution (not Azure Firewall as described above), you must add
packages.aks.azure.com:443
to your allowed egress list by March 31, 2025. After this date, new AKS VHD images will require access to this hostname to successfully provision.Updated Egress rules for AKS clusters are available in the Microsoft Learn documentation.
FAQs
Q. What are the destination IP addresses of the host
packages.aks.azure.com
?You can view the current IP addresses at any time using the following command:
Any changes to the current IP addresses will be communicated at least one week in advance through AKS Release Notes. However, we strongly recommend using a Layer 7 firewall rule that allows traffic to the domain instead of relying on specific IP addresses. This approach avoids the need to update your configuration whenever IP addresses change.
Q. What happens if I do not update my firewall to allow access to the new FQDN by May 31, 2025?
When you perform cluster or Nodepool upgrades, your AKS cluster nodes will use the new Virtual Hard Disk (VHD) image and download necessary binaries from the new FQDN. If you do not update your firewall to allow access to the new FQDN by the full deprecation date, May 31st, 2025, the nodes will fail to be provisioned, and your upgrade operation will fail. However, the existing nodes will continue to function until they are replaced by AKS Node Auto-Healing or a Nodepool upgrade.
Q. How long do I need to allow access to the old FQDN
acs-mirror.azureedge.net:443
?New VHD images that reference the new FQDN-
packages.aks.azure.com
will be released in March 2025. After the new VHD images are applied to all the nodes of your AKS cluster, you can safely remove access to the old FQDN.Q. What happens if my AKS cluster has the node auto-upgrade OS feature enabled, and it updates the cluster nodes to the new VHD that accesses the new FQDN before I add the new FQDN to my firewall rules?
We are implementing a fallback logic in the new VHDs that tries to access the old FQDN if it fails to access the new FQDN-
packages.aks.azure.com
. You might experience a small latency due to this fallback operation, but your node will still successfully provision.Q. What happens if my cluster cannot upgrade to the new VHD image that accesses the new FQDN?
Please note that after May 31,2025, the
acs-mirror.azureedge.net
CDN endpoint will no longer function and therefore any operation that that causes the nodes to be replaced-such as the Node Auto-Healing-will fail to operate. If you are on a legacy Kubernetes cluster, we recommend following our extensive AKS Migration Guide to migrate to a new AKS cluster.Help and Support
If you have questions, get answers from community experts in Microsoft Q&A. If you have a support plan and you need technical help, create a support request.
The text was updated successfully, but these errors were encountered: