eks-long-running-pods: EKS Long Running Pods

日期: 2025-08-01 | 影响软件: eks-long-running-pods | POC: 已公开

漏洞描述

Ensure that Amazon Elastic Kubernetes Service (EKS) clusters do not have pods running for more than 30 days. Long-running pods may indicate stale deployments, potential security risks, or resource inefficiencies.

PoC代码[已公开]

id: eks-long-running-pods

info:
  name: EKS Long Running Pods
  author: princechaddha
  severity: medium
  description: |
    Ensure that Amazon Elastic Kubernetes Service (EKS) clusters do not have pods running for more than 30 days. Long-running pods may indicate stale deployments, potential security risks, or resource inefficiencies.
  impact: |
    Long-running pods can lead to security vulnerabilities, outdated software versions, and inefficient resource utilization. They may also indicate a lack of proper deployment practices and maintenance.
  remediation: |
    Review and update long-running pods. Consider implementing proper deployment strategies, regular updates, and automated pod rotation policies.
  reference:
    - https://kubernetes.io/docs/concepts/workloads/pods/pod-lifecycle/
    - https://docs.aws.amazon.com/eks/latest/userguide/kubernetes-versions.html
  tags: cloud,devops,aws,amazon,eks,aws-cloud-config

variables:
  region: "us-east-1"
  days: "30"

flow: |
  code(1)
  for(let cluster of iterate(template.clusters)){
    set("cluster", cluster)
    code(2)
  }

self-contained: true

code:
  - engine:
      - sh
      - bash
    source: |
      aws eks list-clusters --region $region --query 'clusters' --output json

    extractors:
      - type: json
        name: clusters
        internal: true
        json:
          - '.[]'

  - engine:
      - sh
      - bash
    source: |
      aws eks update-kubeconfig --name $cluster --region $region
      kubectl get pods --all-namespaces --field-selector=status.phase=Running --output=json | \
      jq -r --argjson days "$days" '.items[] |
      select((now - (.metadata.creationTimestamp | fromdate)) > ($days * 24 * 3600)) |
      .metadata.namespace + "/" + .metadata.name'

    matchers:
      - type: word
        words:
          - "/"

    extractors:
      - type: dsl
        dsl:
          - '"EKS cluster " + cluster + " has pods running for more than " + days + " days"'
# digest: 4a0a00473045022078e484dde6ab14a2c60f99ac9498ceac9733bddc7c21db406a696cf49974d53d022100b8fde680e4403da4247605f652a492812f19e6559cfd949652e11e22dadbf28e:922c64590222798bb761d5b6d8e72950