Moonshot — GKE Deployment with Ansible

This project automates the provisioning, deployment, and testing of the Moonshot application on Google Kubernetes Engine (GKE) using Ansible playbooks and roles. It covers the full lifecycle: cluster creation, application deployment with persistent PostgreSQL storage, Horizontal Pod Autoscaling (HPA), load testing with JMeter, and Google Cloud Monitoring dashboards.

---

Goal

To provide a fully automated Infrastructure-as-Code (IaC) solution that:

• Provisions a GKE cluster on Google Cloud Platform;
• Deploys the Moonshot application and its PostgreSQL database on Kubernetes;
• Configures persistent storage and HPA for scalability;
• Validates the deployment through a suite of automated Ansible tests;
• Supports load testing via Apache JMeter;
• Provisions a Google Cloud Monitoring dashboard for observability.

---

Tech Stack

Infrastructure & Orchestration:

• Google Kubernetes Engine (GKE)
• Kubernetes (Deployments, Services, PVCs, HPA)
• Ansible (playbooks + roles)

Application & Database:

• Moonshot (containerised via Docker — hvaz18/ascn:latest)
• PostgreSQL

Load Testing:

• Apache JMeter 5.6.3

Local Development:

• Docker & Docker Compose

Observability:

• Google Cloud Monitoring (custom dashboard via dashboard.json)

---

Prerequisites

• ansible installed locally with the kubernetes.core and google.cloud collections
• gcloud CLI authenticated with a service account JSON key
• kubectl configured (updated automatically during cluster creation)
• Python packages: kubernetes, packaging (installed by the install-k8s role)
• A GCP project with GKE API enabled

---

Configuration

Edit inventory/gcp.yml to match your GCP environment:

gcp_project: <your-gcp-project-id>
gcp_cred_file: ~/your-service-account.json
gcp_zone: us-central1-a
gcp_machine_type: e2-medium
gcp_disk_size_gb: 100
gcp_initial_node_count: 2

Sensitive variables (database credentials, etc.) are stored in group_vars/all.yml and encrypted with Ansible Vault.

---

Deployment Instructions

1. Create the GKE cluster

ansible-playbook gke-cluster-create.yml

2. Deploy Moonshot and all components

# With database seeding
ansible-playbook moonshot-deploy.yml -e seed_database=true

# Without database seeding (preserves existing data)
ansible-playbook moonshot-deploy.yml -e seed_database=false

# Skip HPA deployment
ansible-playbook moonshot-deploy.yml -e skip_hpa=true

3. Undeploy

# Undeploy without deleting data
ansible-playbook moonshot-undeploy.yml -e delete_data=false

# Undeploy and delete all persistent data
ansible-playbook moonshot-undeploy.yml -e delete_data=true

4. Destroy the GKE cluster

ansible-playbook gke-cluster-destroy.yml

---

Testing

Run the full test suite (5 tests covering deployment, DGC creation/access, undeploy, and clean restart):

ansible-playbook test-all.yml

Individual tests can be run using tags:

ansible-playbook test-all.yml --tags test1   # Deploy + accessibility check
ansible-playbook test-all.yml --tags test2   # Digital Green Certificate creation/access
ansible-playbook test-all.yml --tags test3   # Undeploy + confirm inaccessibility
ansible-playbook test-all.yml --tags test4   # Redeploy with preserved data
ansible-playbook test-all.yml --tags test5   # Clean redeploy cycle

---

Load Testing (JMeter)

1. Install JMeter

ansible-playbook installJmeter.yml

2. Run a load test

./runJmeter.sh -i <APP_IP> -t <NUM_THREADS>

Example:

./runJmeter.sh -i 35.222.208.250 -t 50

Results are saved to /opt/jmeter/.../results.txt. The JMeter test plan (MoonshotTesSimpl.jmx) is automatically updated with the provided IP and thread count.

---

Monitoring

Create the Google Cloud Monitoring dashboard:

ansible-playbook createDashboard.yml

The dashboard definition is in dashboard.json.

---

Local Development (Docker)

To run Moonshot locally without Kubernetes:

# Create network
docker network create ascn_tp

# Start PostgreSQL
docker run --name my_postgres \
  -e POSTGRES_DB=mydatabase \
  -e POSTGRES_USER=myuser \
  -e POSTGRES_PASSWORD=mypassword \
  --network ascn_tp \
  -p 5432:5432 \
  -d postgres

# Build and start Moonshot
docker build -t moonshot ./docker
docker run --network ascn_tp -p 8000:8000 -it moonshot

---

Directory Structure

.
├── gke-cluster-create.yml      # Playbook: create GKE cluster
├── gke-cluster-destroy.yml     # Playbook: destroy GKE cluster
├── moonshot-deploy.yml         # Playbook: deploy all components
├── moonshot-undeploy.yml       # Playbook: undeploy all components
├── test-all.yml                # Playbook: run full test suite
├── installJmeter.yml           # Playbook: install JMeter + OpenJDK 19
├── createDashboard.yml         # Playbook: provision GCloud dashboard
├── runJmeter.sh                # Script: run JMeter load tests
├── MoonshotTesSimpl.jmx        # JMeter test plan
├── dashboard.json              # Google Cloud Monitoring dashboard config
├── inventory/
│   └── gcp.yml                 # Inventory + cluster/app variables
├── group_vars/
│   └── all.yml                 # Global variables (Vault-encrypted)
├── docker/
│   ├── Dockerfile              # Moonshot Docker image
│   └── moonshot-main/          # Moonshot application source
└── roles/
    ├── gke_cluster_create/     # GKE cluster provisioning
    ├── gke_cluster_destroy/    # GKE cluster teardown
    ├── install-k8s/            # Python/k8s dependencies
    ├── create-node-dir/        # Node directory setup
    ├── k8s-storage/            # Persistent Volumes provisioning
    ├── postgres-deployment/    # PostgreSQL Deployment
    ├── postgres-pvc/           # PostgreSQL PersistentVolumeClaim
    ├── postgres-service/       # PostgreSQL Service
    ├── moonshot-service/       # Moonshot LoadBalancer Service
    ├── run_moonshot/           # Moonshot Deployment
    ├── deploy_hpa/             # Horizontal Pod Autoscaler
    ├── test_moonshot/          # Test roles (service, DGC)
    └── undeploy-all/           # Teardown all resources

---

Contributors

This project was developed by students from the University of Minho as part of the curricular unit Cloud and Network Applications (ASCN), under the Application Engineering track.

Name	Univ. Id
Gonçalo Marinho	PG55945
Henrique Vaz	PG55947
Mike Pinto	PG55987

---

License

This project is licensed under the MIT License.