Steve Nay

This has been released in NTH app version [`v1.17.0`](https://github.com/aws/aws-node-termination-handler/releases/tag/v1.17.0). Helm chart version [`v0.19.0`](https://github.com/aws/eks-charts/releases/tag/v0.0.100).

Interested to hear from contributors here whether the solution in https://github.com/aws/aws-node-termination-handler/pull/582, which adds the label `node.kubernetes.io/exclude-from-external-load-balancers` to nodes undergoing cordon-and-drain operations, is sufficient for your needs here. Does that solve...

Released in [`v1.17.2`](https://github.com/aws/aws-node-termination-handler/releases/tag/v1.17.2) (chart version `v0.19.2`).

Glad you were able to resolve your issue! Out of curiosity, which version of NTH are you running? Are you using IMDS mode?

@pmcenery-bl Good idea. Can you share the configuration you used for Windows and make a recommendation about how the chart should be different from the linux version?

This has been released in NTH app version [`v1.17.0`](https://github.com/aws/aws-node-termination-handler/releases/tag/v1.17.0). Helm chart version [`v0.19.0`](https://github.com/aws/eks-charts/releases/tag/v0.0.100).

@stevehipwell Tell me if I am understanding your situation correctly: 1. You have a stateful workload that needs to run pods in multiple AZs, balanced in roughly equal ratio across...

Related to #59

I opened a new issue to publish metrics for a similar usecase, in case it's useful to anyone here: https://github.com/aws/containers-roadmap/issues/1587

@sbkg0002 I see a metric for `RepositoryPullCount` in the docs for ECR private (released in January 2022), but not one that describes how recently an image was pulled. Can you...