AWS CloudTrail cost optimization: events, Lake and retention
Published and checked: 12 November 2026Review my CloudTrail costsAWS CloudTrail cost optimization starts by mapping event collection, Lake usage and adjacent logging costs. Review each change against the security evidence your organization needs to preserve.

Map the CloudTrail bill before changing collection
Start with the event route and charging unit. The official CloudTrail pricing page distinguishes trails, Lake, Insights and CloudWatch Logs delivery. The first trail copy of ongoing management events delivered to S3 has no CloudTrail delivery charge; S3 storage and analysis remain separate costs. Additional management copies and data events are charged. Direct delivery to CloudWatch Logs without a trail incurs both CloudTrail delivery and CloudWatch Logs ingestion charges. Separate these routes before comparing costs.
| Cost boundary | Evidence to collect | Decision to review |
|---|---|---|
| Trails | Event category, copies, account and Region | Which collection paths are required? |
| Lake ingestion | Event data store, event source and ingested volume | Which events need this analysis destination? |
| Lake retention and queries | Pricing option, retention and scanned data | Which history and query window serve the use case? |
| Adjacent services | S3 and CloudWatch billing lines for the route | What does the complete logging arrangement cost? |
Ask the security team to identify the authoritative logging path and explain any intentional copies. Add a purpose and owner to each destination. The AWS cost optimization scope guide establishes the wider account boundary. Use the S3 cost review for retained objects and the CloudWatch cost review for that separate part of the bill.
Review data event scope with the security owner
AWS documents that data events record resource operations such as S3 object activity and Lambda invocation. They can be high volume, are not logged by default by trails or event data stores, and incur additional charges. Advanced event selectors can narrow collection to the events required for a use case. Validate the actual configuration rather than inferring coverage from the existence of a trail.
- List the resources and event categories currently selected.
- Identify the detection, investigation or evidence requirement served by each selection.
- Check overlaps between organization and account-level collection.
- Have the security owner approve any proposed exclusion before implementation.
A lower event count is only useful if required evidence remains available. For each proposed selector change, document which investigations would still work and which events would no longer be captured. Ask for a representative validation exercise and a rollback condition. Keep the security requirement beside the financial calculation so the reviewer can evaluate the same proposal.
Review existing Lake ingestion, retention and queries
The CloudTrail Lake cost guidance explains that ingestion uses uncompressed data, while query charges depend on optimized and compressed data scanned. A compressed S3 archive size is therefore a different quantity from the Lake ingestion measure. Record the relevant unit in the estimate before comparing an import with ongoing collection.
- Inventory each event data store and its business owner.
- Record the pricing option, approved retention and collection purpose.
- Separate ongoing ingestion from any historical import.
- Review recurring query windows with the analysts who use the results.
- Compare the full configuration against observed billing before proposing a change.
AWS recommends selecting a Lake pricing option against the event types and expected ingestion, narrowing historical import windows and specifying starting and ending eventTime values in queries. Those measures address different cost drivers. Check the documented Lake controls against the actual store configuration before estimating their effect.
Choose an engagement that hands over an operable result
Request an inventory of collection routes, a reconciliation of the cost boundary and a findings register. The consultant should explain how a proposed change reaches the security approver and who implements it. Ask for evidence of relevant AWS billing and logging experience, and request a permitted sample calculation or walkthrough of the analysis method.
| Deliverable | Acceptance condition |
|---|---|
| Logging inventory | Routes, owners and purposes are visible to the security team. |
| Cost baseline | Usage units and invoice lines reconcile, with gaps explained. |
| Change proposal | Coverage impact, approver and rollback condition are recorded. |
| Handover | Queries, calculations and operating instructions remain usable internally. |
Use the cloud audit pricing checklist to distinguish analysis, tooling and implementation fees. Set the audit access boundary separately from permission to change event collection. If account metadata is sufficient for an initial estimate, agree that narrower evidence set before sharing a broader archive.
Verify the result against the same logging boundary
Keep configuration changes and workload changes visible separately. Compare the accepted billing period with the baseline, explain any growth in application activity and preserve the approved logging requirement. A fall in charges after traffic drops does not establish the effect of a selector change. Ask finance to review the measurement alongside the security acceptance record.
- Record the configuration and approval that were actually implemented.
- Retain the cost calculation with its billing units and exclusions.
- Confirm that the required investigation or evidence check still works.
- Assign ownership for future event growth and query reviews.
For the wider financial follow-up, use the realized savings measurement guide. Keep the logging inventory current when applications, accounts or security requirements change. Ask the proposal to state whether this recurring review belongs to your team or forms a separate scope of work.