AWS Lambda Function High-Frequency Invocation by a Single Principal

IMPORTANT: This documentation is no longer updated. Refer to Elastic's version policy and the latest documentation.

AWS Lambda Function High-Frequency Invocation by a Single Principal

Identifies a single principal directly invoking AWS Lambda functions at a high volume within a one-hour window. Adversaries may drive excessive invocations to abuse functions for resource hijacking or cryptomining, to inflate costs in a denial-of-wallet attack, or to enumerate function behavior. This is a volumetric heuristic: the threshold is environment-dependent and high-throughput applications can exceed it, so tune it to the deployment. This rule relies on AWS Lambda data event logging, which is not enabled by default.

Rule type: esql

Rule indices: None

Severity: medium

Risk score: 47

Runs every: 60m

Searches indices from: now-61m (Date Math format, see also Additional look-back time)

Maximum alerts per execution: 100

References:

Tags:

Domain: Cloud
Data Source: AWS
Data Source: Amazon Web Services
Data Source: AWS CloudTrail
Data Source: AWS Lambda
Use Case: Threat Detection
Tactic: Impact
Resources: Investigation Guide

Version: 2

Rule authors:

Elastic

Rule license: Elastic License v2

Investigation guide

edit

Triage and analysis

Investigating AWS Lambda Function High-Frequency Invocation by a Single Principal

A principal issuing a high volume of direct Lambda invocations in a short window can indicate function abuse for resource hijacking or cryptomining, a denial-of-wallet cost attack, or behavioral enumeration. Because Lambda data events record only the invocation metadata (caller, function, source) and not the function’s internal behavior, this rule is purely volumetric and should be treated as corroborating signal.

Possible investigation steps

Identify the principal in aws.cloudtrail.user_identity.arn and determine whether the volume exceeds its historical baseline.
Determine whether the principal is a known high-throughput application or automation identity, or an unexpected user.
Review source.ip / user_agent.original and recent credential activity for signs of compromise.
Correlate with billing/concurrency metrics and with other Lambda or IAM activity by the same principal.

False positive analysis

High-throughput apps, batch processing, and load tests routinely exceed fixed thresholds. Tune the threshold and exclude known high-volume identities after validation.

Response and remediation

If abuse is confirmed, throttle or disable the affected functions (reserved concurrency), rotate or restrict the principal’s credentials, and review function code and execution-role permissions.
Apply per-function reserved concurrency and account-level guardrails to bound cost and blast radius.

Additional information

Setup

edit

Setup

This rule requires AWS Lambda data events to be logged in CloudTrail and ingested via the AWS integration. Lambda invocation (Invoke) is a data-plane event and is NOT logged by default; enable data event logging for Lambda functions in the trail (optionally scoped to sensitive functions to manage volume). Tune the invocation-count threshold in the query to the environment before enabling.

Rule query

edit

from logs-aws.cloudtrail-*

// Lambda invocation data events (data-plane; requires data event logging enabled)
| where
    event.provider == "lambda.amazonaws.com"
    and event.action like "Invoke*"
    and event.outcome == "success"
    and aws.cloudtrail.user_identity.arn IS NOT NULL

| stats
    Esql.invocation_count = count(*),
    Esql.source_ips = values(source.ip)
  by
    aws.cloudtrail.user_identity.arn

// Threshold is environment-dependent — tune to the deployment
| where Esql.invocation_count >= 1000

| keep
    aws.cloudtrail.user_identity.arn,
    Esql.invocation_count,
    Esql.source_ips

| sort Esql.invocation_count desc

Framework: MITRE ATT&CK^TM

Tactic:
- Name: Impact
- ID: TA0040
- Reference URL: https://attack.mitre.org/tactics/TA0040/
Technique:
- Name: Resource Hijacking
- ID: T1496
- Reference URL: https://attack.mitre.org/techniques/T1496/

« AWS KMS Imported Key Material Deleted AWS IAM User Console Login from Multiple Geolocations »