autopilot

Load testing an application running on an autopilot cluster

This blueprint creates an Autopilot cluster with Google-managed Prometheus enabled and installs an application that scales as the traffic that is hitting the load balancer exposing it grows. It also installs the tooling required to distributed load test with locust on that application and the monitoring tooling required to observe how things evolve in the cluster during the load test. Ansible is used to install the application and all the tooling on a management VM.

The diagram below depicts the architecture.

Running the blueprint

1. Clone this repository or open it in cloud shell, then go through the following steps to create resources:

2. Initialize the terraform configuration

   terraform init
   

3. Apply the terraform configuration

   terraform apply -var project_id=my-project-id
   

4. Copy the IP addresses for grafana, the locust master.

5. Change to the ansible directory and run the following command

   ansible-playbook -v playbook.yaml
   

6. Open to the locust master web interface url in your browser and start the load test

7. SSH to the management VM

   gcloud compute ssh mgmt --project my-project
   

8. Run the following command to check that the application pods are running on different nodes than the load testing and monitoring tooling.

   kubectl get pods -A -o wide
   

9. Run the following command to see how the application pods scale

   kubectl get hpa -n sample -w
   

10. Run the following command to see how the cluster nodes scale

    kubectl get nodes -n
    

Alternatively you can also check all the above using the dashboards available in grafana.

Variables

name	description	type	required	default
project_id	Project ID.	string	✓
cluster_network_config	Cluster network configuration.	object({…})		{…}
deletion_protection	Prevent Terraform from destroying data storage resources (storage buckets, GKE clusters, CloudSQL instances) in this blueprint. When this field is set in Terraform state, a terraform destroy or terraform apply that would delete data storage resources will fail.	bool		false
mgmt_server_config	Management server configuration.	object({…})		{…}
mgmt_subnet_cidr_block	Management subnet IP CIDR range.	string		"10.0.2.0/24"
project_create	Parameters for the creation of the new project.	object({…})		null
region	Region.	string		"europe-west1"
vpc_create	Flag indicating whether the VPC should be created or not.	bool		true
vpc_name	VPC name.	string		"vpc"

Outputs

name	description	sensitive
urls	Grafanam, locust and application URLs.

Test

module "test" {
  source = "./fabric/blueprints/gke/autopilot"
  project_create = {
    billing_account_id = "12345-12345-12345"
    parent             = "folders/123456789"
  }
  project_id = "my-project"
}
# tftest modules=11 resources=36