We analyzed billing report of Indexdata AWS account and found some points where it could be optimized to reduce costs.
- EBS storages in us-west-2 used for Rancher Scratch cluster. (At the moment it is 897+ unmounted EBS Volumes that we can delete: Looks like it is some issue with volumes termination during pods deletion)
- EC2 instances in us-west-2 used for Rancher Scratch cluster. Now all instances of this cluster are on-demand with shape not covered by reserved instances. (Solution: after
RANCHER-12completion move scratch to spot instances or change shape to use reserved instances)
- Carrier-io instance in us-east-1 and us-west-2.
- We should delete old carrier-io instance in us-east-1
- Recreate new carrier-io instance in us-west-2 region with proper shape to use reserved instances