How the attack works
AWS Regions that were launched after 20 March 2019 are disabled in an account until someone opts in. EnableRegion in the AWS Account Management API does that for one Region, and AWS notes that the action propagates the IAM resources of the account, such as users and roles, to the Region. The call is free, and so is the Region until resources run in it. Opting in is asynchronous: AWS documents the states ENABLING, ENABLED, DISABLING and DISABLED.
An organization that works in two or three Regions rarely watches the others. Regional security services may not be switched on there, dashboards and cost reports are filtered to the Regions in use, and people who open the console see nothing unusual in their own Region. AWS's Threat Technique Catalog lists account:EnableRegion under the use of unused cloud Regions for this reason: an attacker who can enable a Region gets a place to run instances or other paid resources that the owner is unlikely to notice until the bill arrives. Enabling several Regions in a row gives more room and more quota.
AWS states that CloudTrail logs all Account Management API operations, under the event source account.amazonaws.com. AWS publishes example records for the contact operations of the same API and none for EnableRegion, so the rules here match the event name only. No incident report was used for this page.
What an attacker needs
- Credentials for a principal allowed account:EnableRegion, which broad administrator policies include.
- No service control policy that denies the action or limits the account to an approved list of Regions.
Impact
The attacker gains Regions in which the organization has no monitoring habits, and uses them to run resources at the owner's cost, for example for compute-heavy workloads, or to keep infrastructure that survives a cleanup of the Regions in use. The change itself harms nothing, which is why it is easy to miss.
Detection rules
Opt-In AWS Region Enabled for an Account
Detects EnableRegion calls to the AWS Account Management API. The call opts an account in to a Region that is disabled by default, which applies to Regions launched after 20 March 2019, and AWS notes that it propagates the IAM resources of the account to that Region. A Region that nobody in the organization watches is a place to run resources unnoticed, and AWS's Threat Technique Catalog lists the event for that reason. AWS states that CloudTrail logs all Account Management API operations under the event source account.amazonaws.com. It publishes no example record for EnableRegion, so the rule matches the event name only: read the Region from the request. It also matches attempts that failed (errorCode present): check errorCode, because an unauthorized attempt is worth seeing.
title: Opt-In AWS Region Enabled for an Account
id: bb22ea3a-d002-443d-bc1c-09a9bf41407e
status: experimental
description: >-
Detects EnableRegion calls to the AWS Account Management API. The call
opts an account in to a Region that is disabled by default, which applies
to Regions launched after 20 March 2019, and AWS notes that it propagates
the IAM resources of the account to that Region. A Region that nobody in
the organization watches is a place to run resources unnoticed, and AWS's
Threat Technique Catalog lists the event for that reason. AWS states that
CloudTrail logs all Account Management API operations under the event
source account.amazonaws.com. It publishes no example record for
EnableRegion, so the rule matches the event name only: read the Region
from the request.
It also matches attempts that failed (errorCode present): check errorCode,
because an unauthorized attempt is worth seeing.
references:
- https://docs.aws.amazon.com/accounts/latest/reference/monitoring-cloudtrail.html
- https://docs.aws.amazon.com/accounts/latest/reference/manage-acct-regions.html
- https://aws.amazon.com/blogs/security/setting-permissions-to-enable-accounts-for-upcoming-aws-regions/
- https://aws-samples.github.io/threat-technique-catalog-for-aws/Techniques/T1535.html
- https://attack.mitre.org/techniques/T1535/
author: Forestall
date: 2026-11-18
tags:
- attack.stealth
- attack.t1535
logsource:
product: aws
service: cloudtrail
detection:
selection:
eventSource: account.amazonaws.com
eventName: EnableRegion
condition: selection
falsepositives:
- A planned expansion into a new Region, done once by an administrator or by the landing zone automation. Confirm the event with the cloud platform team, and add the automation role to a filter_* selection on userIdentity.arn if it enables Regions for new accounts.
level: medium
False positives
- A planned expansion into a new Region, done once by an administrator or by the landing zone automation. Confirm the event with the cloud platform team, and add the automation role to a filter_* selection on userIdentity.arn if it enables Regions for new accounts.
Several Opt-In AWS Regions Enabled by One Principal
Counts EnableRegion calls per calling principal (userIdentity.arn). Each call opts the account in to one Region, so three or more calls by one principal within one hour means that several Regions were switched on in a row, or that the same Region was requested for several accounts. An organization that expands does this one Region at a time and with a plan, while someone who wants as much unwatched room and quota as possible enables everything that is off. Denied attempts are counted as well. The count and the window are starting points to tune. The rule counts calls, not different Regions, because AWS publishes no example record that shows the Region field of this event.
Fires when one userIdentity.arn has at least 3 matching events within 1 hour.
title: Opt-In AWS Region Enabled (Several Regions Base Event)
id: 74c2aa3c-9f84-4acc-8176-72bf9ada667f
name: aws_opt_in_region_enabled
status: experimental
description: >-
Base event of the threshold rule in this file. Any EnableRegion call to
the AWS Account Management API, including attempts that were denied.
logsource:
product: aws
service: cloudtrail
detection:
selection:
eventSource: account.amazonaws.com
eventName: EnableRegion
condition: selection
level: low
---
title: Several Opt-In AWS Regions Enabled by One Principal
id: 45475eea-04b5-4fad-90f4-a106c5610209
status: experimental
description: >-
Counts EnableRegion calls per calling principal (userIdentity.arn). Each
call opts the account in to one Region, so three or more calls by one
principal within one hour means that several Regions were switched on in
a row, or that the same Region was requested for several accounts. An
organization that expands does this one Region at a time and with a plan,
while someone who wants as much unwatched room and quota as possible
enables everything that is off. Denied attempts are counted as well. The
count and the window are starting points to tune. The rule counts calls,
not different Regions, because AWS publishes no example record that shows
the Region field of this event.
references:
- https://docs.aws.amazon.com/accounts/latest/reference/monitoring-cloudtrail.html
- https://docs.aws.amazon.com/accounts/latest/reference/manage-acct-regions.html
- https://aws-samples.github.io/threat-technique-catalog-for-aws/Techniques/T1535.html
- https://attack.mitre.org/techniques/T1535/
author: Forestall
date: 2026-11-18
tags:
- attack.stealth
- attack.t1535
correlation:
type: event_count
rules:
- aws_opt_in_region_enabled
group-by:
- userIdentity.arn
timespan: 1h
condition:
gte: 3
falsepositives:
- Landing zone automation that enables an approved set of Regions for every new account. Exclude that role.
- The values of 3 calls and one hour are starting points, not tested limits.
level: high
False positives
- Landing zone automation that enables an approved set of Regions for every new account. Exclude that role.
- The values of 3 calls and one hour are starting points, not tested limits.
Detection logic
Patterns that need more than one event, a baseline or knowledge of your own environment. Each is a hypothesis to build and tune in your SIEM, not a ready rule.
Resources created in a Region shortly after it was enabled
A planned move into a new Region is followed by a slow, reviewed rollout. An attacker enables a Region in order to use it at once, so compute resources appear there within hours, created by the same principal or by one it made.
- An
EnableRegionevent fromaccount.amazonaws.com. Note the Region of the request (the RegionName parameter of the API) and the caller inuserIdentity.arn. - Within 24 hours, write events whose
awsRegionis that Region, for exampleRunInstancesfromec2.amazonaws.com. - Alert when the creator is the principal that enabled the Region or a principal that was created after that call.
EventsEnableRegionRunInstances
LimitsA real expansion into a new Region produces the same sequence, usually by a deployment pipeline. The 24 hours are a starting point. The Region parameter follows the API reference, and AWS publishes no example record for the event, so confirm the field in your own events first.
Write activity in a Region your organization does not use
Regions that were launched before the opt-in model are always enabled, so no event announces their use. Any change in a Region that is not on the organization's list is worth a look, whoever made it.
- Keep a list of the Regions in which your organization deploys.
- Collect management events that are not read-only (
readOnlyis false) and readawsRegion. - Alert when the Region is not on the list, and leave out global services, which log in one fixed Region.
LimitsIt needs a maintained list of Regions and exceptions for global services such as IAM. AWS services that act on your behalf can write events in Regions you never chose.
Triage
- Identify the caller (userIdentity.arn), the source IP address and the user agent. Did anyone plan to start working in a new Region?
- Read the Region of the request and check whether more Regions were enabled by the same principal around the same time.
- Look for the first activity in the new Region after the event, with priority on new instances, functions, containers and service quota requests.
- Check what else the principal did in the account before and after the call.
Response
Once the attack is confirmed:
- Contain the principal that enabled the Region, and review everything else it did.
- Inventory the resources in the newly enabled Regions and remove what nobody in the organization created.
- Disable the Region again once it is empty. AWS notes that disabling a Region removes all IAM access to resources that reside in it, so clean up first.
- Put a service control policy in place that denies the Region opt-in actions and limits usable Regions.
Mitigation
- Deny account:EnableRegion and account:DisableRegion with a service control policy for everyone except a small number of administrators.
- Limit the Regions in which principals may act with a service control policy on the
aws:RequestedRegioncondition key, so that an enabled Region is still unusable. - Use a trail that applies to all Regions, so that activity in a newly enabled Region is logged without further setup.
- Enable your security services and your cost alerts for all Regions, not only for the ones in use.
Frequently asked questions
Which Regions are affected?
AWS's account actions for enabling and disabling Regions apply to Regions launched after 20 March 2019. Those are disabled in an account by default and need an opt-in. Regions that existed before that date are always enabled, so activity in an older Region that you do not use produces no `EnableRegion` event and has to be found by watching the Region field of all events.
Does enabling a Region cost anything or change existing resources?
No. AWS states that enabling or disabling a Region is free and that you are charged only for resources you create in the new Region. The opt-in propagates the account's IAM resources to the Region. Nothing else in the account changes, which is why the event is the only early sign.
Will my trail record what happens in the new Region?
A trail that applies to all Regions logs events from every Region of the partition, and AWS documents that a trail created in the console does so by default. A trail that was limited to one Region shows only the events of that Region. Check which kind you have before you rely on seeing activity in a Region that was just enabled.
How often does this happen legitimately?
Rarely. An organization opts in to a Region when it starts to deploy there, which is a planned decision that the platform team knows about. One event from an administrator with a known project is normal. Several Regions enabled in a short time by one principal is not.
References
- MITRE ATT&CK T1535: Unused/Unsupported Cloud Regions
- Logging AWS Account Management API calls using AWS CloudTrail
- Enable or disable AWS Regions in your account
- AWS Security Blog: Setting permissions to enable accounts for upcoming AWS Regions
- Threat Technique Catalog for AWS: T1535 Unused/Unsupported Cloud Regions