Redshift clusters should use enhanced VPC routing

How to fix it

• Can break clients: COPY and UNLOAD fail unless the VPC routes them, through an S3 gateway endpoint for same-Region buckets or a NAT gateway otherwise, and IAM Identity Center sign-in needs interface endpoints.
• Interrupts service: Redshift restarts the cluster automatically when this setting changes, so schedule the change for a quiet period.

In the console

Redshift → Clusters → the cluster → Properties → Network and security settings → Edit → Enhanced VPC routing

Enhanced VPC routing

Turn on, once the VPC has a network path for COPY and UNLOAD

With the AWS CLI

aws redshift modify-cluster --cluster-identifier <cluster-id> --enhanced-vpc-routing --region <region>

With Terraform

resource "aws_redshift_cluster" "this" {
  cluster_identifier   = var.cluster_identifier
  node_type            = var.node_type
  enhanced_vpc_routing = true
}