Share via

AKS deployment memory pressure - Node pool memory working set

Ammar Mohanna 26 Reputation points
2022-12-02T14:42:23.917+00:00
  • Description:
    I created an AKS cluster, and I am adding several services to it (around 10), each one of these is configured to have 3 pods.

The node pool has the following parameter:
Node pools: 1 node pool
Kubernetes versions: 1.23.12
Node sizes: Standard_DS2_v2

266675-monitoring.png

  • Status quo:
    Some of the pods are stuck on 'Pending' status, and therefore their API response remains '503 Service Temporarily Unavailable'.
    And sometimes, the pods are turning off and on constantly.
  • Questions:
    Am I using too many services for this size of the cluster?
    Are there any suggestions or additional information that I can provide to make my question clearer?
Azure Kubernetes Service
Azure Kubernetes Service

An Azure service that provides serverless Kubernetes, an integrated continuous integration and continuous delivery experience, and enterprise-grade security and governance.

0 comments
Answer accepted by question author
  1. Lj-MSFT 81 Reputation points Microsoft Employee
    2022-12-06T10:23:00.04+00:00

    @Ammar Mohanna : yes, it seems that there is memory pressure on the nodes. To validate the same, you can connect to your AKS cluster and run kubectl top nodes, which will show you the CPU and memory usage on the nodes, and will help you to confirm if there is with lack of memory, and if yes, you can scale up the number of nodes to be able to get those pods into running state which are currently stuck in pending.

    1 person found this answer helpful.

0 additional answers

Sort by: Most helpful
Most helpful Newest Oldest

Your answer

Answers can be marked as 'Accepted' by the question author and 'Recommended' by moderators, which helps users know the answer solved the author's problem.