Why is my AWS Config bill so high?
AWS Config charges per configuration item recorded, per rule evaluation and per conformance pack evaluation, in each account and Region. A configuration item is recorded every time a recorded resource or one of its relationships changes, so spikes almost always come from high-churn resources: autoscaling instances, containers creating network interfaces, or automation changing security groups and tags many times a day.
How the charge is calculated
| Meter | Example price (US East (N. Virginia), pricing page) |
|---|---|
| Configuration item, continuous recording | $0.003 each |
| Configuration item, periodic (daily) recording | $0.012 each, at most one per resource per day when it changed |
| Rule evaluation | $0.001 each for the first 100,000, cheaper above |
| Conformance pack evaluation | $0.001 each for the first 100,000, cheaper above |
Prices are from the AWS Config pricing page examples at the time of writing. A resource that changes ten times a day costs ten continuous items, or one periodic item. Relationship changes count too: attaching a network interface records items for the instance, the interface and the security group.
Find what is generating items
- Cost Explorer: Service = Config, group by Usage type and Region. Separate item costs from rule evaluation costs.
- In the AWS Config dashboard, the Configuration Items Recorded usage metric can be split by resource type. The same metrics are in CloudWatch under the
AWS/Confignamespace. - Look for resource types that churn:
AWS::EC2::NetworkInterface,AWS::EC2::Instancein autoscaling groups,AWS::EC2::SecurityGroupedited by automation, and resources tagged by pipelines on every run.
# Which Config metrics are available in this Region
aws cloudwatch list-metrics --namespace AWS/Config --query 'Metrics[].[MetricName,Dimensions[0].Value]' --output text
# Current recorder settings: what is recorded and how often
aws configservice describe-configuration-recordersReduce it, without losing what compliance needs
- Record high-churn types daily instead of continuously, with recording mode overrides.
- Exclude resource types nobody evaluates, especially in non-production accounts.
- Fix the churn: pipelines that re-tag or rewrite security groups on every run, or very short-lived instances.
- Trim rules and conformance packs to what you act on, and remove duplicates deployed both per account and through the organisation.
{
"name": "default",
"roleARN": "arn:aws:iam::123456789012:role/aws-service-role/config.amazonaws.com/AWSServiceRoleForConfig",
"recordingGroup": {
"allSupported": false,
"exclusionByResourceTypes": {"resourceTypes": ["AWS::EC2::NetworkInterface"]},
"recordingStrategy": {"useOnly": "EXCLUSION_BY_RESOURCE_TYPES"}
},
"recordingMode": {
"recordingFrequency": "CONTINUOUS",
"recordingModeOverrides": [{
"description": "daily for churny types",
"resourceTypes": ["AWS::EC2::Instance", "AWS::EC2::SecurityGroup"],
"recordingFrequency": "DAILY"
}]
}
}Apply with aws configservice put-configuration-recorder --configuration-recorder file://recorder.json, using your existing recorder name and role.
What not to do
- Do not stop the recorder in accounts covered by audit or security requirements; agree exclusions with whoever owns compliance.
- Do not exclude types that your Config rules or Security Hub controls evaluate; those findings go stale silently.
- Do not assume daily recording is always cheaper: a resource that changes once a day costs more as a periodic item than as a continuous one.
Questions
Why did my AWS Config cost spike?
Usually a resource type started changing much more often, for example autoscaling, container network interfaces, or automation rewriting tags or security groups.
Is AWS Config periodic recording cheaper?
For resources that change several times a day, yes: you pay at most one periodic item per day. For rarely changing resources, continuous can be cheaper.
Does stopping the Config recorder delete history?
No. You can stop recording and still access previously recorded configuration items.
Config costs track churn, so they often spike the week a new autoscaling or container workload ships. CloudConsole, a phone app for AWS, shows spend by service and usage type, so a Config spike is visible on your phone the day it starts.
About the app → · Download on the App Store