π€ Welcome to PCF Migration Agent!
I help you migrate applications from Pivotal Cloud Foundry (PCF) to any Kubernetes platform:
βοΈ GKE, EKS, AKS, OKE, or π΄ OpenShift. This guide explains each step and answers common questions.
π Migration Steps
π 1. Repository Analysis
Analyzes your application's source code from Git to understand its structure, dependencies, and configuration.
Detects the technology stack and creates an initial assessment.
π 2. PCF Live Analysis
Connects to your PCF platform to gather live application data including configuration, environment variables,
service bindings, and resource requirements. Essential for PCF-to-GKE migrations.
βοΈ 3. Transform
Converts your application into Kubernetes manifests (Deployments, Services, ConfigMaps) and generates a Dockerfile.
Uses Move2Kube for cloud-agnostic transformation.
ποΈ 4. Stage
Builds your application into a Docker container image and pushes it to your target platform's registry
(Artifact Registry, ECR, ACR, OCIR, or OpenShift Registry). Creates platform-specific CI/CD configuration.
π 5. Deploy
Deploys your containerized application to your selected Kubernetes platform (GKE, EKS, AKS, OKE, or OpenShift).
Applies all Kubernetes manifests and verifies the deployment is successful.
π 6. Generate Docs
Creates comprehensive migration documentation including guides, architecture diagrams,
comparison reports, and operational runbooks.
β Frequently Asked Questions
What is PCF to Kubernetes migration?
It's the process of moving applications from Pivotal Cloud Foundry (PCF) to Kubernetes (GKE).
This involves converting PCF-specific configurations to Kubernetes manifests, containerizing applications,
and adapting service bindings and environment variables.
Do I need source code or can I migrate from live PCF?
Both options are supported! You can either:
- Git Repository: Provide a Git URL for source code analysis
- PCF Live: Connect to your PCF platform to migrate running applications without source code
What buildpacks are supported?
We support all major PCF buildpacks:
- Java (Spring Boot, Maven, Gradle)
- Node.js
- Python
- Go
- .NET Core
- PHP
- Static Files (NGINX)
- Binary
How are credentials and secrets handled?
Security is a priority:
- GCP Secret Manager: Store credentials securely in Google Cloud
- Runtime Fetching: Secrets are fetched at runtime, never stored in code
- Encrypted Storage: All sensitive data is encrypted
- No Logging: Credentials are never logged or exposed
What if a step fails?
If a step fails:
- Check the Activity Log tab for detailed error information
- Review the error message for specific suggestions
- Fix the issue (e.g., correct credentials, fix configuration)
- Re-run the failed step - you don't need to start over
Each step is independent and can be retried without affecting completed steps.
How long does a complete migration take?
Typical timeline for a complete migration:
- Analyze: 30-60 seconds
- PCF Analysis: 20-40 seconds
- Transform: 1-2 minutes
- Stage: 2-5 minutes (depends on app size)
- Deploy: 1-3 minutes
- Generate Docs: 30-60 seconds
Total: Approximately 5-12 minutes for a complete migration.
Can I customize the generated Kubernetes manifests?
Yes! After the Transform step completes:
- Download the generated manifests from the Reports tab
- Edit the YAML files to customize resources, limits, or configurations
- Apply them manually using
kubectl apply -f
The generated manifests are a starting point and can be fully customized for your needs.
π‘ Tips & Best Practices
- Start with PCF Analysis: For live apps, run PCF Analysis first to capture runtime configuration
- Review Activity Logs: Check the Activity tab after each step to understand what happened
- Use GCP Secrets: Store credentials in Secret Manager for better security
- Test in Dev First: Deploy to a development cluster before production
- Generate Documentation: Always run Generate Docs for a complete migration record