How tigera-solutions/cc-gke-demo works
Project type and purpose
• This is a Kubernetes PoC/demo repository for Calico Cloud on Google Kubernetes Engine (GKE).
• It automates provisioning a GKE cluster, deploying the Online Boutique microservices app, and applying Calico Cloud network‐security and observability features.
File structure
• Top level:
• manifests/ – Calico network policy YAMLs organized into three tiers (application, platform, security).
• scripts/ – Shell scripts that provision the cluster, deploy the sample app, onboard to Calico Cloud, create tiers and policies, then clean up.
• docs/ – Quickstart, prerequisites, troubleshooting, observability guidance, FAQ, references and cheatsheet.
• validation/ – Test scenarios and validation script for end‐to‐end checks.
• No application source code here: the Online Boutique microservices manifests are applied by scripts.
Key insights from README
• Purpose: “Zero to secured app in under an hour.”
• Steps: cluster provisioning → app deployment → Calico Cloud onboarding → apply network policies → observe and validate → teardown.
• Demonstrates: microservices traffic flows, security enforcement, live observability.
Main components
a. User and CLI:
• Operator’s workstation running the scripts (gcloud, kubectl, calicoctl).
b. GCP Environment:
• Google Cloud account and VPC (network).
• GKE control plane (managed).
• GKE worker nodes in a node pool.
c. Calico Cloud:
• Calico Cloud management plane (hosted by Tigera).
• Calico Cloud agent (calico-node DaemonSet) running on each GKE node.
d. Online Boutique microservices:
• Frontend service (LoadBalancer).
• Internal microservices: productcatalog, currency, recommendation, paymentservice, loadgen, redis, etc.
e. External dependencies:
• GitHub (for egress).
• DNS provider (cluster DNS).
f. Observability components:
• Metrics/Logs exported to Calico Cloud or Prometheus/Grafana (if configured by scripts).
g. Network policy tiers:
• Tier-security, platform, application (enforced by Calico CNI).
Relationships and data flows
• Operator invokes scripts → uses gcloud to create GKE cluster → uses kubectl to deploy pods → Calico agent registers with Calico Cloud.
• End users → HTTP(S) → Frontend LoadBalancer → Frontend Pod.
• Frontend Pod → gRPC/HTTP → productcatalog, currency, recommendation, paymentservice Pods.
• loadgen Pod generates traffic for demonstration.
• Pods interact with external services only where allowed by Calico policies (e.g., egress to GitHub, DNS).
• Calico agent enforces policies at the pod‐to‐pod network level.
• Observability data flows from agents to Calico Cloud management plane.
Architectural patterns and design principles
• Microservices architecture: many small services communicating over the network.
• Layered network security: separate tiers for default‐deny policies, platform policies (e.g., allow kube-api, DNS), application policies (allow only needed service-to-service flows).
• Infrastructure-as-code: scripts automate provisioning, deployment, policy application, teardown.
• Zero-trust networking: default deny, explicit egress/ingress rules.
Essential details to capture
• Clear separation of responsibilities: operator vs GKE vs Calico Cloud.
• Explicit policy tiers and how they map onto pod communication flows.
• Observability and validation components.