Skip to main content
compliance · aws

Accounts where the AWS Config recorder is switched off

resource types
1
rule IDs covered
1
severity
medium

What does ZopNight detect here?

ZopNight emits one config-recorder finding per AWS account, firing only when the discoverer confirms the recorder is disabled. Coverage is currently probed in us-east-1 alone, so multi-region accounts are judged from that 1 region: a deliberate, documented limit rather than a silent assumption about the other regions.

Signal and threshold

How ZopNight evaluates Accounts where the AWS Config recorder is switched off.
Field Value
Rule IDsRC-1520
Categorycompliance
Severitymedium
Metricnone — pure configuration read
Sourceconfig_recorder.go

What you lose without the recorder

AWS Config is the service that remembers what your infrastructure looked like yesterday. With the recorder off there is no configuration timeline: when an incident review asks “when did this security group open port 22, and what else changed that day”, the honest answer becomes “we don’t know”. Config Rules, conformance packs, and drift detection all sit on top of the recorder. None of them function in an account where it never ran.

How the signal is produced

The discoverer emits a single synthetic config-recorder resource per account and stamps a boolean config_recorder_enabled on it. The rule fires only when that flag is present and false. If the flag is missing or unreadable (a failed probe, a permissions gap), the rule abstains instead of declaring an unknown account non-compliant. An earlier version read a tag that no producer ever wrote and therefore never fired at all; the current gate reads the field the producer actually populates.

The single-region caveat, stated plainly

The probe currently runs against us-east-1 only, and one resource represents the whole account. CIS AWS Foundations treats Config as an all-regions control, so an account with the recorder on in us-east-1 but off everywhere else will pass this rule today. That accuracy bound lives in the producer, not the rule. The fail-closed gate is correct for whatever scope the probe reports, and widening coverage means aggregating recorder state across regions upstream, not loosening the check.

Verify from your terminal

Terminal window
aws configservice describe-configuration-recorder-status --region us-east-1 \
--query 'ConfigurationRecordersStatus[].[name,recording,lastStatus]' --output table

An empty result means no recorder exists; recording: false means one exists but is stopped.

Turning it on without a billing surprise

Config bills per configuration item recorded, so the pragmatic setup is to record all resource types but check the estimate for high-churn accounts first. An account cycling hundreds of spot instances daily records a lot of items. Enable it once through the console or IaC, point snapshots at a versioned S3 bucket, and the finding clears at the next discovery pass. The recommendation itself carries no savings figure; the value is forensic, not financial.

See it fire on your bill.

Connect an account read-only. The first findings land in minutes.

417 rule families across 353 resource types on 22 platforms. Every threshold, metric, and IAM action is documented on these pages before you grant anything.

417 rule families documented
353 resource types covered
read-only default access level
Multi-cloud automation· Production-ready in 30 min· SOC 2 · ISO 27001· 20–60% off the bill, first month· 4 platforms · 1 console· Multi-cloud automation· Production-ready in 30 min· SOC 2 · ISO 27001· 20–60% off the bill, first month· 4 platforms · 1 console·