Skip to content

GCP Google Cloud Platform - Create log-based Alert policy

Step-01: Introduction

  • Unusual Request Methods: Monitor for requests using uncommon HTTP methods such as PUT, DELETE, or CONNECT, which may indicate attempts to exploit less common vulnerabilities or misconfigurations.
  • Create a log-based alert when we find HTTP methods PUT, DELETE, CONNECT in Nginx Acess logs
  • Review the incidents in Alerting

Step-02: Generate HTTP Traffic for HTTP methods PUT, DELETE, CONNECT

# HTTP METHODS simulation using curl
curl -X PUT http://34.27.233.153
curl -X DELETE http://34.27.233.153
curl -X CONNECT http://34.27.233.153

# Generate Traffic in a while loop in Cloud shell
while true; do curl -X PUT http://34.27.233.153; sleep 1; done
while true; do curl -X DELETE http://34.27.233.153; sleep 1; done
while true; do curl -X CONNECT http://34.27.233.153; sleep 1; done

Step-03: Logs Explorer: Search for PUT, DELETE, CONNECT HTTP METHODS

# Search for HTTP Methods PUT, CONNECT, DELETE in nginx access logs
resource.type="gce_instance"
log_id("nginx_access")
httpRequest.requestMethod="PUT" OR httpRequest.requestMethod="DELETE" OR httpRequest.requestMethod="CONNECT"
labels."compute.googleapis.com/resource_name"="myvm1"

Step-03: Create Log based Alert policy

  • Go to -> Cloud Logging / Cloud Monitoring -> Logs Explorer
  • Create Alert from Logs Explorer screen
  • Click on Create Alert

Alert Details

  • Alert Policy Name: suspicious-http-methods
  • Policy Severity Level: Critical
  • Documentation: leave empty
  • Click on NEXT

Choose logs to include in the alert

  • Define log entries to alert on:
    # Search for HTTP Methods PUT, CONNECT, DELETE in nginx access logs
    resource.type="gce_instance"
    log_id("nginx_access")
    httpRequest.requestMethod="PUT" OR httpRequest.requestMethod="DELETE" OR httpRequest.requestMethod="CONNECT"
    labels."compute.googleapis.com/resource_name"="myvm1"
    
  • Click on NEXT

Set notification frequency and autoclose duration

  • Time between notifications: 5 min
  • Incident autoclose duration: 7 days
  • Click on NEXT

Who should be notified? (optional)

  • Notiication Channels: gcpuser08
  • Click on SAVE

Step-04: Verify Incidents

  • Go to Cloud Logging or Cloud Monitoring -> Detect -> Alerting
  • Verify Incidents
  • Verify email id for sending incidents
🎉 New Course
Ultimate DevOps Real-World Project Implementation on AWS
$15.99 $84.99 81% OFF
DEVOPS2026FEB
Enroll Now on Udemy
🎉 Offer