Detection: Baseline of Excessive AWS Instances Terminated by User - MLTK

REMOVED DETECTION

This detection has been removed from the Splunk Threat Research content library and is no longer maintained or supported.

Reason: All detection(s) which leverage this baseline have been deprecated. As such, this baseline has been deprecated as well.

Removed in version: 5.2.0

If you have any questions or concerns, please reach out to us at research@splunk.com.

Updated Date: 2026-05-13 ID: b28ed6de-e4ba-40f7-ae0a-93a088c774ab Author: Jason Brewer, Splunk Type: Baseline Product: Splunk Enterprise Security

Description

This search is used to build a Machine Learning Toolkit (MLTK) model for how many TerminateInstances users do in the environment. By default, the search uses the last 90 days of data to build the model. The model created by this search is then used in the corresponding detection search, which identifies subsequent outliers in the number of TerminateInstances performed by a user in a small time window.

Search

1`cloudtrail` eventName=TerminateInstances errorCode=success 
2| bucket span=10m _time 
3| stats count as instances_terminated by _time src_user 
4| fit DensityFunction instances_terminated threshold=0.0005 into ec2_excessive_terminateinstances_v1

Data Source

No data sources specified for this detection.

Macros Used

Name	Value
cloudtrail	`sourcetype=aws:cloudtrail`

Annotations

- MITRE ATT&CK

+ Kill Chain Phases

+ NIST

+ CIS

- Threat Actors

ID	Technique	Tactic

Default Configuration

This detection is configured by default in Splunk Enterprise Security to run with the following settings:

Setting	Value
Disabled	true
Cron Schedule	`N/A`
Earliest Time	`N/A`
Latest Time	`N/A`
Schedule Window	`N/A`
Creates Finding (Notable)	No
Creates Intermediate Finding (Risk Event)	No

Implementation

You must install the AWS App for Splunk (version 5.1.0 or later) and Splunk Add-on for AWS (version 4.4.0 or later), then configure your CloudTrail inputs. In addition, you must have the Machine Learning Toolkit (MLTK) version >= 4.2 installed, along with any required dependencies. Depending on the number of users in your environment, you may also need to adjust the value for max_inputs in the MLTK settings for the DensityFunction algorithm, then ensure that the search completes in a reasonable timeframe. By default, the search builds the model using the past 30 days of data. You can modify the search window to build the model over a longer period of time, which may give you better results. You may also want to periodically re-run this search to rebuild the model with the latest data. More information on the algorithm used in the search can be found at https://docs.splunk.com/Documentation/MLApp/4.2.0/User/Algorithms#DensityFunction.

Known False Positives

none

Associated Analytic Story

Suspicious AWS EC2 Activities

Detection Testing

Test Type	Status	Dataset	Source	Sourcetype
Validation	Not Applicable	N/A	N/A	N/A
Unit	Not Applicable	N/A	`N/A`	`N/A`
Integration	Not Applicable	N/A	`N/A`	`N/A`

Replay any dataset to Splunk Enterprise by using our replay.py tool or the UI. Alternatively you can replay a dataset into a Splunk Attack Range

Source: GitHub |

Version: 1