Azure-docs: AKS Cluster Autoscaler pod logs

[amadav]

I am trying to analyze the scaling activity that AKS cluster can do with cluster autoscaler enabled. For this, I am looking into how long a node takes to come up and get registered with the masters and alike for scale down scenario. For this, I have been traditionally using the cluster-autoscaler pod logs and look for logs pertaining to scaleUp and scaleDown events. I would appreciate if I could get a handle on the same when I enable cluster-autoscaler on an AKS cluster. Currently using the preview feature for CAS from here - https://docs.microsoft.com/en-us/azure/aks/cluster-autoscaler. I am okay just knowing the VMSS activity logs too, I suppose. Would appreciate any pointers.

---

Document Details

⚠ Do not edit this section. It is required for docs.microsoft.com ➟ GitHub issue linking.

• ID: 077f1bb5-2ba9-490c-2c6e-0de292266539
• Version Independent ID: 0fd87b4f-88a4-c2eb-3ebc-f2b113ac203c
• Content: Use the cluster autoscaler in Azure Kubernetes Service (AKS)
• Content Source: articles/aks/cluster-autoscaler.md
• Service: container-service
• GitHub Login: @mlearned
• Microsoft Alias: mlearned

[Alberto-Vega]

@amadav Thanks for the feedback! We are currently investigating and will update you shortly.

[jakaruna-MSFT]

@amadav Are you looking for a place or a command to get those events?

@jluk @mlearned Please add your comments

[jakaruna-MSFT]

@amadav is there any update

[jakaruna-MSFT]

I will close this out for now. If you need additional help please let me know and we can reopen and continue.

[amadav]

When autoscaler is enabled using the "az aks" command, I am not able to get the logs for the autoscaler possibly because it is not run as part of agent-pool nodes. Is there a way to get/monitor the logs on the master components?

[jemag]

@jakaruna-MSFT I am also trying to find any kind of logs to debug the autoscaler. Despite having 10 pods "Pending" because of lack of resources, the cluster is still not scaling up and I have no logs to debug it.

[amadav]

@jakaruna-MSFT , can you please reopen the issue?

[ohorvath]

@jakaruna-MSFT please reopen this one, up-down scaling events are really crucial. We have to use kubernetes cluster autoscaler instead of the built-in one, because of this issue.

[jluk]

@amadav and @ohorvath I've submitted a PR with instructions on how to retrieve the autoscaler logs and status. Look to this document in the next 2 days for an update. https://docs.microsoft.com/en-us/azure/aks/cluster-autoscaler

In the meantime you can retrieve logs by following this doc and checking the cluster-autoscaler box.
https://docs.microsoft.com/azure/aks/view-master-logs#enable-diagnostics-logs

You can retrieve CA health by reading the configmap with this cmd:

kubectl get configmap -n kube-system cluster-autoscaler-status -o yaml