[gardener-extension-provider-aws]

⚠️ Breaking Changes

[USER] The gp2 StorageClass is now removed. (gardener/gardener-extension-provider-aws#576, @StenlyTU)
[OPERATOR] This version of provider-aws requires Gardener v1.50.0. (gardener/gardener-extension-provider-aws#562, @ialidzhikov)
[OPERATOR] This version of provider-aws requires Gardener v1.50+. (gardener/gardener-extension-provider-aws#571, @kris94)

[OPERATOR] CloudProfileConfig now supports a new field .machineImages[].versions[].regions[].architecture. It specifies the supported CPU architecture of the given machine image AMI. (gardener/gardener-extension-provider-aws#565, @acumino)
[OPERATOR] WorkerStatus now supports a new field .machineImage[].architecture. It specifies the supported CPU architecture of the given worker pool. (gardener/gardener-extension-provider-aws#565, @acumino)

[USER] Users can now set IOPS for a GP3 volume type. Validation of IOPS (i.e. whether it is allowed and is in the specified range for a volume type) is done on the AWS side, so feedback will arrive once the volume is created. (gardener/gardener-extension-provider-aws#561, @rishabh-11)
[OPERATOR] provider-aws now mutates the cluster-autoscaler Deployment by implementing the EnsureClusterAutoscalerDeployment function. This is required in the context of kubernetes/autoscaler#4517 - cluster-autoscaler supports --feature-gates flag and provider extensions have to mutate the cluster-autoscaler Deployment to add the CSI related feature gates to it. (gardener/gardener-extension-provider-aws#562, @ialidzhikov)

[USER] The following images are updated: (gardener/gardener-extension-provider-aws#540, @acumino)
- k8s.gcr.io/sig-storage/csi-provisioner: v2.1.2 -> v2.2.2 (for kubernetes < 1.20)
- k8s.gcr.io/sig-storage/csi-provisioner: v2.1.2 -> v3.2.0 (for kubernetes >= 1.20)
- k8s.gcr.io/sig-storage/csi-attacher: v3.3.0 -> v3.4.0
- k8s.gcr.io/sig-storage/csi-resizer: v0.5.0 -> v1.5.0
- k8s.gcr.io/sig-storage/csi-snapshotter: v3.0.3 -> v4.2.1 (for kubernetes >= 1.20)
- k8s.gcr.io/sig-storage/snapshot-validation-webhook: v3.0.3 -> v4.2.1 (for kubernetes >= 1.20)
- k8s.gcr.io/sig-storage/snapshot-controller: v3.0.3 -> v4.2.1 (for kubernetes >= 1.20)
- k8s.gcr.io/sig-storage/csi-node-driver-registrar: v1.3.0 -> v2.5.1
- k8s.gcr.io/sig-storage/livenessprobe: v2.3.0 -> v2.7.0
[USER] The following image is updated: (gardener/gardener-extension-provider-aws#574, @ialidzhikov)
- k8s.gcr.io/provider-aws/aws-ebs-csi-driver: v1.5.3 -> v1.9.0
[OPERATOR] The extension now uses distroless instead of alpine as a base image. (gardener/gardener-extension-provider-aws#564, @dimityrmirchev)
[OPERATOR] The following dependency is updated: (gardener/gardener-extension-provider-aws#568, @acumino)
- github.com/gardener/gardener: v1.48.0 -> v1.50.0
[OPERATOR] Update golang version used to 1.18 (gardener/gardener-extension-provider-aws#569, @kon-angelo)
[OPERATOR] Update MTU-resizer alpine image (gardener/gardener-extension-provider-aws#579, @kon-angelo)

[OPERATOR] Use go mod instead of dep (gardener/aws-lb-readvertiser#21, @kon-angelo)
- Update golang version used.
[OPERATOR] The aws-lb-readvertiser now uses distroless instead of alpine as a base image. (gardener/aws-lb-readvertiser#23, @dimityrmirchev)

[OPERATOR] The default leader election resource lock of machine-controller-manager has been changed from endpointsleases to leases. (gardener/machine-controller-manager#711, @acumino)
- Please make sure, that you had at least machine-controller-manager@v0.43.0 running before upgrading to v0.46.0, so that it has successfully acquired leadership with the hybrid resource lock (endpointsleases) at least once.

[USER] Rollout freeze won't happen due to Unknown machines now. (gardener/machine-controller-manager#733, @himanshu-kun)

[OPERATOR] Published docker images for Machine-Controller-Manager are now multi-arch ready. They support linux/amd64 and linux/arm64. (gardener/machine-controller-manager#732, @timuthy)
[OPERATOR] The machine-controller-manager container now uses distroless instead of alpine as a base image. (gardener/machine-controller-manager#734, @dimityrmirchev)

[OPERATOR] machine-controller-manager-provider-aws now uses distroless instead of alpine as a base image. (gardener/machine-controller-manager-provider-aws#90, @ialidzhikov)
[DEVELOPER] probeResources() now doesn't try to delete orphan resources but only lists them. (gardener/machine-controller-manager-provider-aws#85, @Mkmittal)
- The beforeSuite for IT test now calls for cleanup of orphan resources separately.
- The Integration Test, which looks for orphan resources, now doesn't try to delete the orphan resources and just waits for them to be done automatically.

[OPERATOR] Terraform google provider is updated to v4.19.0 (gardener/terraformer#119, @bd3lage)