AKS agentpool scaling is stuck for over 4 hours. Node count remains 9 instead of 10, with no error shown.

Question

AKS agentpool scaling is stuck for over 4 hours. Node count remains 9 instead of 10, with no error shown.

Muzammil - Unicorn lab 0

AKS agentpool scaling has been stuck for more than 4 hours. The target node count is 10, but the current node count remains 9. The cluster operation status shows Succeeded, while the agentpool provisioning state remains Scaling. Activity Log shows no errors. VMSS capacity is still 9, and the Scale node pool option is disabled. Please help identify and resolve the stuck scaling operation.

Sujitha Lakkireddy 85 Reputation points Microsoft External Staff Moderator

2026-07-16T17:16:20.35+00:00

Hello Muzammil - Unicorn lab,

Thank you for choosing Microsoft Q&A forum.

Thanks for sharing the details. Since the cluster operation status is Succeeded while the node pool provisioning state remains Scaling, and the VMSS capacity is still 9, this could indicate that the scale-out request is stuck at the AKS or VMSS layer.

As initial troubleshooting, please check:

az aks nodepool show \

--resource-group <resource-group> \

--cluster-name <aks-cluster> \

--name <nodepool-name>

az vmss show \

--resource-group <managed-resource-group> \

--name <vmss-name> Also verify:

Activity Logs for both the AKS cluster and the underlying VMSS.

Regional quota and SKU availability:

az vm list-usage --location <region>

Resource locks:

az lock list --resource-group <resource-group>

If Cluster Autoscaler is enabled, whether another scaling operation is in progress.

If the issue persists, please share the output of az aks nodepool show, the node pool provisioning state, and any recent scaling, upgrade, or configuration changes so we can investigate further.
Muzammil - Unicorn lab 0 Reputation points

2026-07-16T19:04:10.67+00:00

I retried the node pool scaling, but it has been stuck in the Scaling state for over 4 hours again with no progress. The previous attempt failed with a FailedToCreateOrUpdateVMSS internal error.
Muzammil - Unicorn lab 0 Reputation points

2026-07-16T19:05:13.49+00:00

Thank you for your response.

I have completed the requested checks.

AKS node pool (agentpool) is still in Scaling state.

VMSS ProvisioningState = Failed.

VMSS capacity remains 9 although the requested node count is 10.

Cluster operation status is Succeeded.

Activity Log shows no errors.

Regional quota is available (Standard BS Family: 30/33 vCPUs).

No resource locks are configured.

No recent changes were made except scaling the node pool from 9 to 10.

The scaling operation has now been stuck for more than 4 hours. Could you please advise how to recover the stuck scaling operation or identify the underlying VMSS failure?

1 answer

Your answer

Sujitha Lakkireddy 85 Reputation points Microsoft External Staff Moderator

2026-07-16T17:16:20.35+00:00

Hello Muzammil - Unicorn lab,

Thank you for choosing Microsoft Q&A forum.

Thanks for sharing the details. Since the cluster operation status is Succeeded while the node pool provisioning state remains Scaling, and the VMSS capacity is still 9, this could indicate that the scale-out request is stuck at the AKS or VMSS layer.

As initial troubleshooting, please check:

az aks nodepool show \

--resource-group <resource-group> \

--cluster-name <aks-cluster> \

--name <nodepool-name>

az vmss show \

--resource-group <managed-resource-group> \

--name <vmss-name> Also verify:

Activity Logs for both the AKS cluster and the underlying VMSS.

Regional quota and SKU availability:

az vm list-usage --location <region>

Resource locks:

az lock list --resource-group <resource-group>

If Cluster Autoscaler is enabled, whether another scaling operation is in progress.

If the issue persists, please share the output of az aks nodepool show, the node pool provisioning state, and any recent scaling, upgrade, or configuration changes so we can investigate further.
Muzammil - Unicorn lab 0 Reputation points

2026-07-16T19:04:10.67+00:00

I retried the node pool scaling, but it has been stuck in the Scaling state for over 4 hours again with no progress. The previous attempt failed with a FailedToCreateOrUpdateVMSS internal error.
Muzammil - Unicorn lab 0 Reputation points

2026-07-16T19:05:13.49+00:00

Thank you for your response.

I have completed the requested checks.

AKS node pool (agentpool) is still in Scaling state.

VMSS ProvisioningState = Failed.

VMSS capacity remains 9 although the requested node count is 10.

Cluster operation status is Succeeded.

Activity Log shows no errors.

Regional quota is available (Standard BS Family: 30/33 vCPUs).

No resource locks are configured.

No recent changes were made except scaling the node pool from 9 to 10.

The scaling operation has now been stuck for more than 4 hours. Could you please advise how to recover the stuck scaling operation or identify the underlying VMSS failure?

Answer 1

The important detail is VMSS ProvisioningState = Failed. The AKS node-pool “Scaling” state is a symptom: the tenth VM was not created, even though the cluster-level operation reports Succeeded.

First, inspect the Activity Log in the AKS node resource group, not only the cluster resource group. You can get its name from the cluster’s nodeResourceGroup property. Scope the log to the VM scale set and open the failed write operation’s JSON details; that usually exposes capacity, networking, image, policy, or allocation errors hidden by the generic FailedToCreateOrUpdateVMSS message.

Because the node-pool operation has remained nonterminal for hours, abort the last node-pool operation using Azure CLI 2.47 or later:

az aks nodepool operation-abort -g <resource-group> --cluster-name <cluster> -n <nodepool>

Wait until the node pool becomes Canceled, then retry the AKS node-pool scale operation. Aborting releases the lock but does not roll back resources, so verify the pool and VMSS state before retrying. Do not manually change the VMSS capacity; AKS owns that scale set.

Also, 30/33 vCPUs only leaves three vCPUs of family quota, so confirm the selected node size fits that remaining quota and that the SKU has regional/zonal capacity. If the retry again returns FailedToCreateOrUpdateVMSS, open an AKS support case with the cluster resource ID, node-pool name, last operation ID, and VMSS deployment correlation ID; that requires backend investigation.