Node autoscaler is not working.

sai theja 0 Reputation points
2023-03-26T05:16:05.4933333+00:00

Hi Team,

We had a recent issue, where node autoscaler is not successful,

We kept the node to be in autoscale mode and quota of the nodes to the specific region are available.

And no taints are provided onto nodes.

Yet, one of the pod was in pending state and no node is allocated to the pod.

On describe of the pending pod, it states, insufficient CPU and memory on one of the node.

As per my understanding if node is not available to allocate, it has to create a new node for the allocation of pod and run.

But it didn't happened, later the memory config of pod is reduced it got allocated, and started running.

Please help me on what might be the issue and node is not getting autoscaled.

Details:

Node count : 17

Node range : 40

Node config: Standard DS12V2.

Regional quota is available.

max pods per node is 30.

pod config:

Requests:

CPU: 2

memory : 14 GB.

Later changed to 10 GB and it started running and no new node allocation is done.

Azure Kubernetes Service (AKS)
Azure Kubernetes Service (AKS)
An Azure service that provides serverless Kubernetes, an integrated continuous integration and continuous delivery experience, and enterprise-grade security and governance.
1,855 questions
{count} votes