karpenter-provider-aws

karpenter-provider-aws copied to clipboard

Avoid subnets that don't have available IP Addresses

11

### Description **What problem are you trying to solve?** When a subnet is almost out of IPs, Karpenter will continue to launch nodes in it, leading to the VPC CNI...

ellistarn

Show runtime defaults in Status field of the EC2NodeClass

1

**Tell us about your request** Some fields in the EC2NodeClass spec (specifically the AWS Cloud Provider), are runtime defaulted, meaning that Karpenter assumes a default value at runtime. Some examples...

bwagner5

Karpenter not honoring/working with topology spread constraints or pod affinity

1

### Description **Observed Behavior**: This will be rather long as I describe the different scenarios we've tested. We have a deployment that, by default, does not specify `topologySpreadConstraints` or `affinity`...

bwmetcalf

Metric karpenter_interruption_received_messages(message_type="SpotInterruptionKind") is not accurate

7

### Description **Observed Behavior**: Currently Karpenter recieves the account all the spot interruption and filters in the karpenter controller logic. But we have setup a lambda to filter the interruption...

hitsub2

Feature request: More flexible requirements for NodePools

1

### Description **What problem are you trying to solve?** I have a generic `NodePool` that's used by a majority of my workloads, which has pretty lose requirements, except that it...

mrparkers

Karpenter on Fargate Profile Results in Failed Readiness/Liveness Checks

6

### Description **Observed Behavior**: When a Fargate Profile is created for our `karpenter` namespace, the pods are restarted to move from the Managed Node Group to Fargate. However, they always...

kgochenour

Karpenter disrespecting Pod Disruption Budget while downscaling of pod

1

### Description **Observed Behavior**: When sudden load are input on deployment, pod got scaled up to 10 pods on new nodes and existing nodes but during downscaling of pods it...

coniferrkishor0874

vmMemoryOverheadPercent design is causing over provision

7

### Description I created a new issue instead of Re-opening https://github.com/aws/karpenter-provider-aws/issues/6611 As I don't have a permission to re-open it. The reason I want to re-open it is documented in...

bnetzi

Karpenter metrics documentation should be improved

2

### Description **How can the docs be improved?** For examples, the following metrics documentation should be improved. 1.karpenter_nodes_termination_time_seconds Examples ```shell karpenter_nodes_termination_time_seconds{app_kubernetes_io_instance="karpenter", app_kubernetes_io_managed_by="Helm", app_kubernetes_io_name="karpenter", app_kubernetes_io_version="0.36.0", helm_sh_chart="karpenter-0.36.0", instance="100.64.81.229:8000", job="kubernetes-service-endpoints", kubernetes_namespace="karpenter", nodepool="celeborn-worker-graviton", quantile="0.9",...

hitsub2

during drift, karpenter destroy old new before new node is ready

2

### Description **Observed Behavior**: when changing instance-category, karpenter will spin up a node but delete the old new before the new node is ready. at times i've even seen it...

christophercorn