AWS Events and Changes To Track For Your AWS Account Security
AWS CloudTrail service captures and publishes every action taken or change performed across any other AWS services.
As of 2026 — “Every action” needs a qualification. CloudTrail Event History shows management events for the last 90 days by default. Data events (such as S3 object reads and Lambda invocations) are opt-in and billed separately, and keeping history beyond 90 days requires a trail. See viewing CloudTrail events and CloudTrail concepts. For configuration advice, see best practices for AWS CloudTrail.
These changes could come as a result of
- Console use by a user
- An application running its code
- A SaaS partner using assume role running/ accessing some API in your account
- An event trigger another component (e.g File got uploaded on S3 which trigger a Lambda function)
As you can see everything gets pushed into CloudTrail. Identity-related events like these gain meaning only when correlated with the surrounding posture, which is the lesson behind correlating identity with misconfigurations so contextual severity reflects the real IAM graph.
The example below shows that the IAM user Alice used the AWS CLI to call the CreateUser action to create a new user named Bob.
{
"Records": [
{
"eventVersion": "1.0",
"userIdentity": {
"type": "IAMUser",
"principalId": "EX_PRINCIPAL_ID",
"arn": "arn:aws:iam::123456789012:user/Alice",
"accountId": "123456789012",
"accessKeyId": "EXAMPLE_KEY_ID",
"userName": "Alice"
},
"eventTime": "2014-03-24T21:11:59Z",
"eventSource": "iam.amazonaws.com",
"eventName": "CreateUser",
"awsRegion": "us-east-2",
"sourceIPAddress": "127.0.0.1",
"userAgent": "aws-cli/1.3.2 Python/2.7.5 Windows/7",
"requestParameters": { "userName": "Bob" },
"responseElements": {
"user": {
"createDate": "Mar 24, 2014 9:11:59 PM",
"userName": "Bob",
"arn": "arn:aws:iam::123456789012:user/Bob",
"path": "/",
"userId": "EXAMPLEUSERID"
}
}
}
]
}
As of 2026 — The sample record above is from the AWS documentation circa 2014 (
eventVersion1.0, an old CLI user agent). Current records use much newer event versions with additional fields, so build detections against the current CloudTrail record contents reference.
So much data?
Data -> information -> insights becomes challenging and overwhelming if the amount is large and needs to be consumed in real-time. Real-time monitoring paired with contextual severity is what keeps this manageable, scoring each event by actual risk instead of drowning your team in noise. So, below is the list of metrics that we recommend you keep an eye on. This doesn’t mean that other events are not important but the objective of this blog is to ensure that the signal-to-noise ratio for your team is minimum and you do not lose sight of critical updates.
ConsoleLogin
- What? This event is raised when anyone logs into the account using the AWS console.
- Why? Generally, if you have CI/ CD setup and your team uses command-line tools to interact with cloud API, then console login is a rare act. It makes it important to stay alert if there is a console login.
- Risk: Medium
- What to do next? If you have a console login, especially for someone with a root or administrator role, then checking with them offline (over Slack or other messenger-based tools) is recommended.
As of 2026 — The assumption that console logins are rare no longer holds for most teams, because IAM Identity Center and SSO sign-ins are routine. Alert on the combinations that matter: any root user sign-in (treat as High risk), and console sign-ins where
additionalEventData.MFAUsedisNo. Use the identity, source IP and device context to decide what is unusual.
StopLogging
- What? When AuditTrail logs are disabled. Yes, it raises this event and logs this action too 🙂
- Why? There will be little to no reason that anyone without a planned notice and communication has disabled AuditTrail logs themselves. So, this event itself is a good enough reason to jump on it.
- Risk: High
- What to do next? Jump on a call with team members (if required), login into your console, and assess who did it? If it was a mistake, then reenable the logging again. Also, put controls and education in a place where this mistake doesn’t happen again in the future. Events like this are exactly why a continuous cloud audit trail matters, and why teams invest in closing the audit-trail gap from missing logs to complete access auditing.
As of 2026 — “AuditTrail” here means AWS CloudTrail. There is no separate AuditTrail product. Also watch for related events that disable or weaken logging:
DeleteTrail,UpdateTrailandPutEventSelectors. Use a service control policy and an organization trail so member accounts cannot switch logging off.
CreateRoute, CreateNetworkAclEntry, CreateSecurityGroup
- What? These deal with VPC security – one of the building blocks in which your compute and other resources and data resides.
- Why? This activity should be planned, documented, and approved activity.
- Risk: High
- What to do next? Alert your team about this and fix the issue/ process which has caused this.
AuthorizeDBSecurityGroupIngress, CreateDBSecurityGroup, DeleteDBSecurityGroup, RevokeDBSecurityGroupIngress
- What? These events deal with database security groups, which control network access to your databases.
- Why? Your database should be one of the most protected resources. Not only shouldn’t it have a public IP assigned, but it also should be in a private subnet. Listening for these events ensures that any changes to DB security groups don’t go unnoticed.
- Risk: High
- What to do next? Alert your team about this and verify that the change was planned and approved.
As of 2026 — DB security groups are an EC2-Classic feature, and EC2-Classic was retired on 15 August 2022, so these four events no longer apply to current RDS databases. Databases in a VPC use ordinary VPC security groups. The events to watch instead are
AuthorizeSecurityGroupIngressandRevokeSecurityGroupIngress(especially rules open to0.0.0.0/0),ModifyDBInstance(changes toPubliclyAccessible) andModifyDBSnapshotAttributeorModifyDBClusterSnapshotAttribute(sharing snapshots). See complete list of AWS RDS misconfigurations.
References
As of 2026: What has changed
Beyond the events above, security teams commonly alert on the following CloudTrail events.
- Logging and detection tampering:
DeleteTrail,UpdateTrail,PutEventSelectors,DeleteFlowLogs,StopConfigurationRecorder,DeleteDetector(GuardDuty). - Identity and access:
CreateAccessKey,CreateLoginProfile,AttachUserPolicy,PutUserPolicy,AttachRolePolicy, unusual cross-accountAssumeRole, and any root user activity. - Data exposure:
PutBucketPolicy,DeletePublicAccessBlock,PutBucketAcl, public snapshot sharing. - Encryption:
ScheduleKeyDeletion,DisableKey.
Rather than reading logs by hand, route these through Amazon EventBridge rules, or use GuardDuty and Security Hub. Note that CloudTrail Lake is no longer open to new customers from 31 May 2026, per AWS’s notice, so plan on a trail with S3 and Athena, or a SIEM. For a managed approach, see cloud detection and response and AWS CSPM implementation guide.
People Also Read
- AWS CSPM Implementation Guide
- Complete List of AWS RDS Misconfigurations
- What is Cloud Detection and Response?
- AWS Real-Time Security Events
- Cloudanix Query Builder: Write Queries On Your Data
- All You Need To Know About AWS CloudTrail
- Best Practices For AWS CloudTrail
- CloudTrail, CloudWatch, Splunk, Custom logs solution
- A Practical Step by Step Guide To Prevent Security Breach Of Your AWS Account
- A Big List Of Mistakes AWS Cloud Users Have Done And Spent Millions
- How to setup Multi Factor Authentication (MFA) devices in AWS?
- The Different Types of Protection your Application Needs If Hosted in AWS
- Top 18 Challenges of Cloud Security in 2026