Loading

AWS Bedrock Provisioned Model Throughput Tampering

Detects creation, modification, or deletion of AWS Bedrock Provisioned Model Throughput via the CreateProvisionedModelThroughput, UpdateProvisionedModelThroughput, and DeleteProvisionedModelThroughput APIs. Provisioned Throughput reserves dedicated, billed model capacity for Amazon Bedrock. An adversary who scales this capacity up can drive large, unauthorized cost (cloud resource/bill hijacking), while deleting reserved throughput can cause denial of service to production workloads that depend on that committed capacity. These control-plane changes should be validated against approved capacity-planning and change-management processes.

Rule type: query
Rule indices:

  • logs-aws.cloudtrail-*

Rule Severity: medium
Risk Score: 47
Runs every:
Searches indices from: now-6m
Maximum alerts per execution: 100
References:

Tags:

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

Version: 1
Rule authors:

  • Elastic

Rule license: Elastic License v2

Amazon Bedrock Provisioned Throughput reserves dedicated, billed model capacity for foundation models. Because this capacity is committed and metered, adversaries can abuse it in two ways: scaling capacity up to incur large, unauthorized cloud spend (resource/bill hijacking), or deleting reserved throughput to deny service to production workloads that rely on committed capacity. This rule identifies CreateProvisionedModelThroughput, UpdateProvisionedModelThroughput, and DeleteProvisionedModelThroughput calls so responders can confirm whether the change was authorized.

  • Identify the actor and context
    • Review aws.cloudtrail.user_identity.arn, aws.cloudtrail.user_identity.type, aws.cloudtrail.user_identity.access_key_id, user_agent.original, and source.ip.
    • Determine whether the identity is an approved administrator, ML/platform engineer, or automation role.
    • Confirm a corresponding change request or capacity-planning ticket exists.
  • Validate the request details
    • Inspect aws.cloudtrail.request_parameters and aws.cloudtrail.response_elements for the model ID, commitment duration, and requested model units. Unusually large model-unit counts or long commitment terms on a Create/Update may indicate cost-driven abuse.
    • For DeleteProvisionedModelThroughput, identify which provisioned model was removed and whether any production workload depended on it.
  • Correlate activity
    • Review other Bedrock control-plane actions (e.g., model invocation logging changes, guardrail changes) and IAM/STS activity from the same identity around the same time.
    • Check cloud.account.id and cloud.region for whether the activity occurred in an expected account/region.
  • Capacity planning: Platform, ML, or FinOps teams may legitimately create, update, or delete provisioned throughput. Validate against change tickets and standard capacity-management procedures.
  • Automation: IaC or deployment pipelines may manage provisioned throughput on bootstrap or teardown. Confirm the source IP and ARN match expected automation infrastructure.
  • If unauthorized, immediately disable the offending access key or role and revert the change (delete unauthorized provisioned throughput, or recreate deleted reserved capacity required by production).
  • Review billing and Cost Explorer for unexpected Bedrock provisioned-throughput charges.
  • Audit the actor's recent activity and rotate credentials if compromise is suspected.
  • Restrict bedrock:CreateProvisionedModelThroughput, bedrock:UpdateProvisionedModelThroughput, and bedrock:DeleteProvisionedModelThroughput to a small set of administrative roles and enforce approval workflows and budget alarms.
data_stream.dataset: "aws.cloudtrail" and
    event.provider: "bedrock.amazonaws.com" and
    event.action: (
        "CreateProvisionedModelThroughput" or
        "UpdateProvisionedModelThroughput" or
        "DeleteProvisionedModelThroughput"
    ) and
    event.outcome: "success"

Framework: MITRE ATT&CK