Jens-Uwe Walther

Adopt Helm chart to mimic Amazon EKS CoreDNS naming conventions

2

- modified Helm chart to mimic Amazon EKS CoreDNS deployment - minor changes due to PDB - modified values.yaml resources to not contain hyphen in name because this is not...

Update Grafana Capacity and Performance dashboards to use new v1 API metrics

2

Fixes [#6767](https://github.com/aws/karpenter-provider-aws/issues/6767) **Description** This PR implements the new v1 metrics according to v1 migration [Updated metrics](https://karpenter.sh/preview/upgrading/v1-migration/#updated-metrics) in sample Grafana Capacity and Performance dashboards. **How was this change tested?** EKS 1.30...

### Description **Observed Behavior**: Current Grafana dashboard used in [Monitoring with Grafana (optional)](https://karpenter.sh/docs/getting-started/getting-started-with-karpenter/#monitoring-with-grafana-optional) does not contain the changes in V1 metrics mentioned in **v1 migration** [Updated metrics](https://karpenter.sh/preview/upgrading/v1-migration/#updated-metrics) **Expected Behavior**: Grafana...

preStop hook should account for AWS NLB deregistration delay

1

Currently during HPA scale-in of NGINX Ingress pods sometimes are terminated faster then the AWS NLB target deregistration delay `deregistration_delay.timeout_seconds`, which leads to traffic send to non-existing targets/pods and client...

Turn off returning HTTP OK for /healthz endpoint during preStop hook to allow AWS NLB unhealthy draining

2

AWS NLB has a feature called unhealthy draining, see [Connection termination for unhealthy targets](https://docs.aws.amazon.com/elasticloadbalancing/latest/network/target-group-health.html#unhealthy-target-connection-termination) which together with the additional feature _connection termination disabled_ allows that the target is considered unhealthy...