Parsivex scans your AWS account for 26 types of waste across compute, storage, database, and networking services. Each finding type below links to a dedicated article explaining what triggers the flag, how to fix it, and typical savings ranges.
Titles and short descriptions match what you see on finding cards in your scan reports.
Compute
| Finding type | What it means |
|---|---|
| Idle EC2 instance | An EC2 instance with very low CPU utilization that can be stopped or terminated. |
| Idle Lambda function | A Lambda function with zero invocations in the lookback window, still incurring storage charges. |
| Over-provisioned Lambda memory | A Lambda function allocated more memory than its invocation metrics justify. |
| Oversized EC2 instance | An EC2 instance running on a larger instance type than its workload requires. |
| Previous-generation EC2 instance type | An EC2 instance running on an older instance family with a cheaper, modern equivalent available. |
| Reserved Instance opportunity | Stable compute usage that could save money with Reserved Instance or Savings Plan pricing. |
Storage
| Finding type | What it means |
|---|---|
| CloudWatch Logs with no retention policy | A CloudWatch log group set to never expire, allowing unbounded log storage charges to accumulate. |
| EBS gp2 volume not migrated to gp3 | An EBS gp2 volume that could be migrated to gp3 for lower cost and better baseline performance. |
| EFS file system missing lifecycle policy | An EFS file system with no lifecycle policy to transition infrequently accessed data to Infrequent Access. |
| Idle EFS file system | An EFS file system with no or near-zero read/write activity still incurring storage charges. |
| Incomplete multipart uploads | Aborted S3 multipart uploads still consuming storage with no lifecycle rule to clean them up. |
| Orphaned EBS snapshot | An EBS snapshot with no associated volume or AMI still incurring storage charges. |
| Orphaned EBS volume | An unattached EBS volume still incurring storage charges. |
| S3 bucket missing lifecycle policy | A versioned or growing S3 bucket with no lifecycle rules configured to manage object storage costs. |
| S3 wrong storage class | An S3 bucket without lifecycle rules to transition infrequently accessed objects to cheaper tiers. |
Database
| Finding type | What it means |
|---|---|
| Idle ElastiCache cluster | An ElastiCache cluster with near-zero cache traffic or connections still incurring node charges. |
| Idle RDS instance | An RDS database instance with zero or near-zero database connections over the lookback window. |
| Idle Redshift cluster | A Redshift cluster with no active queries or very low utilization still incurring compute charges. |
| Orphaned RDS manual snapshot | An RDS manual snapshot whose source database instance no longer exists, still incurring storage charges. |
| Over-provisioned DynamoDB capacity | A DynamoDB table on provisioned billing whose consumed capacity is a small fraction of what is provisioned. |
| Over-provisioned ElastiCache node type | An ElastiCache cluster running on a larger node type than its workload metrics justify. |
| Over-provisioned RDS instance | An RDS database running on a larger instance class than its workload requires. |
Networking
| Finding type | What it means |
|---|---|
| Idle VPC interface endpoint | A VPC interface endpoint with near-zero traffic still incurring hourly availability charges. |
| NAT Gateway overuse | High NAT Gateway data-processing charges that could be reduced with VPC endpoints. |
| Unused Elastic IP | An allocated Elastic IP address not associated with a running instance. |
Compute & networking
| Finding type | What it means |
|---|---|
| Idle load balancer | A load balancer with no healthy targets or near-zero traffic still incurring hourly charges. |