A Parsivex scan is a read-only analysis of your AWS account. It pulls billing, inventory, and usage data, checks your resources for waste, enriches findings with plain-English explanations, and produces a savings report. This article explains what happens during a scan.
What data Parsivex collects
Parsivex reads three categories of AWS data during every scan:
| Data source | What we collect | Used for |
|---|---|---|
| AWS Cost Explorer | Monthly and daily spend by service, reservation utilization, Savings Plans coverage | Cost-based findings (NAT Gateway overuse, Reserved Instance opportunities) and report context |
| Resource inventory | EC2 instances, EBS volumes and snapshots, Elastic IPs, RDS databases, ElastiCache clusters, Redshift clusters, Lambda functions, S3 buckets, Elastic Load Balancers, DynamoDB tables | Identifying unused, idle, or oversized resources |
| CloudWatch metrics | CPU, network, request counts, and related utilization signals over a trailing window | Distinguishing idle resources from actively used ones |
All access is read-only through your cross-account IAM role. Parsivex never modifies AWS resources during a scan.
Parsivex does not read S3 object contents, database rows, application logs, or secrets. See What data can Parsivex see? for the full breakdown.
What happens during a scan
After you click Start Scan, Parsivex starts the scan in the background and runs these steps in order:
Cost Explorer data
Fetches account-level billing and cost breakdowns once per scan. If Cost Explorer is unavailable (not enabled or access denied), the scan continues with empty cost data and records a partial scan warning.
Inventory collection
Lists resources in each selected region: EC2, RDS, ElastiCache, Redshift, Lambda, S3, EBS, Elastic IPs, load balancers, and DynamoDB tables. Per-resource API errors (for example, access denied on a single bucket) are logged and skipped — they do not fail the scan.
Metrics collection
Pulls CloudWatch utilization metrics for running EC2 instances, active RDS databases, available ElastiCache and Redshift clusters, invoked Lambda functions, provisioned DynamoDB tables, and load balancers in each selected region. Missing metrics for one resource do not stop the scan.
Waste detection
Checks for specific waste patterns — idle EC2, orphaned EBS volumes, oversized RDS, and more. Cost Explorer-based findings run once per scan; regional detectors run in each selected region. A single check failing is logged but does not fail the scan.
Plain-English explanations
Adds clear descriptions and remediation guidance to each finding.
Report generation
Sorts findings by estimated savings, computes totals, and saves your report. If warnings were recorded, they appear on the report and scan results page.
The progress screen groups these stages into three user-facing phases:
| Progress | Phase label | What is happening |
|---|---|---|
| 0–33% | Collecting data | Inventory, metrics, and Cost Explorer |
| 34–66% | Analysing | Waste detection and remediation script generation |
| 67–100% | Building report | Report assembly and recommendations |
How long scans take
Most scans finish in a few minutes. Duration depends on how many resources you have and how responsive AWS APIs are at the time. Large accounts with hundreds of EC2 instances or many S3 buckets may take longer because metrics and bucket checks run per resource.
You can leave the progress page and come back — the scan continues in the background. When it finishes, your dashboard and scan history update automatically.
What "complete" means
A scan is complete when:
- All scan steps finish without a fatal error
- The scan finishes with 100% progress
- A savings report is generated and linked from your dashboard
A complete scan may still have partial scan warnings if some data sources were unavailable. Warnings mean coverage was incomplete — not that the scan failed.
A scan is failed when a required step cannot finish (for example, inventory collection for a core service like EC2 or RDS returns a hard error). Failed scans show an error message and a Try Again button. They do not generate a new report and do not mark previous findings as resolved. See Finding lifecycle for how resolution works.
Regions scanned
Parsivex covers your account at two scopes:
| Scope | What is scanned |
|---|---|
| Account-wide | Cost Explorer billing data (spend across all AWS regions where you have usage). These findings appear under Global / Account-wide in your report. |
| Regional | EC2, RDS, Lambda, ELB, DynamoDB, EBS, and CloudWatch metrics in each region included in the scan |
Choosing regions
Select one or more AWS regions (or Select all regions) for each scan. Parsivex runs regional inventory, metrics, and detectors once per selected region, merges the results into one report, and shows per-region progress while the scan runs. A multi-region scan still counts as one scan toward your monthly limit. Re-scans from the dashboard, command palette, or Run scan button reuse your saved region selection.
To change the default regions after onboarding, open Integrations → AWS connections and click Edit regions on the connection card. See Verify your AWS connection for where regions are first chosen during setup.
If one region fails (for example, the IAM role cannot read EC2 in ap-south-1), Parsivex marks that region as failed, continues other regions, and still completes the scan when at least one region succeeds.
If you are unsure which regions to include, check the AWS Console region selector or look at where your production EC2 and RDS instances run. Cost Explorer data is always account-wide regardless of how many regions you select.
After the scan
Once complete, you can:
- View your savings report with findings sorted by estimated monthly savings
- Dismiss or snooze findings that are intentional — see Finding lifecycle
- Run remediation scripts on paid plans — see Remediation script safety
Related articles
- Partial scan warnings — Why scans can be incomplete and how to fix warnings
- Severity and savings estimates — How findings are prioritized and savings are calculated
- Your first scan — Step-by-step walkthrough from dashboard to report
- Connection troubleshooting — Fix IAM permission issues that cause partial scans