These are the current recommendations for your clouds. Select any recommendation to see details or alternative options.
Alternatively, select an object type (e.g. instances, volumes, etc) in the navigation bar to the left to get more details per recommendation.
If an IP address is attached to a stopped instance, or not attached to an instance, you will be charged extra.
Cloud Optimizer will recommend which IP addresses should be released.
Instances are charged by type and region. Instances are often over-sized or under-sized. Even right-sized instances can be moved to other type families to save money.
Cloud Optimizer will make recommendations for right-sizing and optimizing cost based on CPU and memory utilization.
If CloudWatch memory utilization monitoring is not enabled, Cloud Optimizer will provide a 'safety' rating and suggest enabling memory utilization monitoring.
GPU Instances are charged by type and region. GPU Instances are often over-sized or under-sized. Even right-sized GPU instances can be moved to other type families to save money.
Cloud Optimizer will make recommendations for right-sizing and optimizing cost based on CPU, GPU, and memory utilization.
If CloudWatch GPU monitoring is not enabled, Cloud Optimizer will provide a ‘safety’ rating and suggest enabling GPU utilization monitoring.
If CloudWatch memory utilization monitoring is not enabled, Cloud Optimizer will provide a ‘safety’ rating and suggest enabling memory utilization monitoring.
Load balancers are charged by time as well as load (requests, bandwidth, etc.).
RDS instances are charged by type and region. Instances are often over-sized or under-sized. Even right-sized instances can be moved to other type families to save money.
Cloud Optimizer will make recommendations for right-sizing and optimizing cost based on CPU and memory utilization.
RDS storage is charged based on volume type and allocated size.
Cloud Optimizer can suggest an optimal size based on disk usage.
Snapshots are an efficient way to back up the entire contents of a disk volume for point-in-time recovery. Snapshots are charged based on changed blocks from the previous snapshot.
Too few snapshots may make it difficult to restore to a particular time. Too many snapshots are unnecessarily expensive.
Cloud Optimizer will flag orphan snapshots and snapshots that do not adhere to retention policies.
Snapshots are an efficient way to back up the entire contents of a disk volume for point-in-time recovery. Snapshots are charged based on changed blocks from the previous snapshot.
Too few snapshots may make it difficult to restore to a particular time. Too many snapshots are unnecessarily expensive.
Cloud Optimizer will flag orphan snapshots and snapshots that do not adhere to retention policies.
Volumes are charged based on volume type and allocated size.
Orphan volumes are usually not needed and can be deleted (with or without archive).
If CloudWatch Disk Monitoring is enabled, Cloud Optimizer can suggest an optimal size based on disk usage.
Workspaces are charged based on running mode.
Workspaces not logged into after the number of days set in the policy workspace:days should be shut down.
Workspaces used less than 80 hours a month should use a running mode of AUTO_STOP.
Workspaces used more than 80 hours a month should use a running mode of ALWAYS_ON.