How softwareag/webmethods-api-gateway works
Project Type and Purpose
• This repo is not a single microservice or UI project, but a comprehensive “resource and sample” repository for the webMethods API Gateway product.
• It contains:
• API definitions and Swagger JSON for each gateway service (apigatewayservices/)
• Postman collections for testing those APIs
• Architecture and operations documentation (docs/)
• Getting-started examples (gettingstarted/)
• Deployment samples for Docker, Kubernetes, Helm, OpenShift, and Command Central (samples/)
• Purpose: illustrate how to deploy, configure, extend, and interact with the API Gateway runtime and its supporting components.
File Structure Analysis
• Top-level directories:
• apigatewayservices – full list of gateway service Swagger definitions + Postman collections
• docs – deep dives into architecture (e.g. Elasticsearch/Kibana setup, security), operations, policies, diagnostics, features
• gettingstarted – minimal GraphQL and Kubernetes examples for a quick launch
• samples – end-to-end deployment templates for Docker, Kubernetes (plain manifests, Helm charts), OpenShift, Command Central
• attachments – static images and zip archives used by the README
• Patterns:
• Control plane vs data plane separation (management APIs, UI docs vs runtime service definitions)
• Infrastructure-as-code (Helm, YAML manifests)
• Observability integration (Elasticsearch + Kibana)
Key Insights from README
• API Gateway provides:
• A web UI for admin and API design
• A runtime cluster that applies policies and routes incoming API traffic
• Analytics dashboard backed by Elasticsearch/Kibana
• DevOps options: Cloud, Docker image, CLI/Command-Central, Helm charts
• Users import APIs via archives, assign policies, create applications (API keys), and monitor analytics.
Main Components
A. Control Plane
1) Admin UI (web application served by gateway runtime)
2) Management REST APIs (Swagger definitions under apigatewayservices/)
3) DevOps Integration: Command Central / CLI / Helm charts / Docker Compose
B. Data Plane
1) API Gateway Runtime Cluster (Nginx ingress or built-in HTTP listener)
2) Policy Enforcement Engine (global policies, security, transformation)
3) Routing to Backend Services (customer-owned microservices, SOAP/REST endpoints)
C. Data Stores & Caches
1) Configuration & state DB (internal data store)
2) Analytics Store: Elasticsearch cluster
3) Caching Tier: Terracotta cluster (session and policy cache)
D. Observability
1) Kibana UI
2) Metrics & transaction logs (shipped from gateway via Elasticsearch plugins)
E. External Systems
1) API Portal (for developer onboarding)
2) Backend services and external APIs
3) Identity provider (OAuth/OpenID Connect)
Interactions and Data Flows
• Management flow (Control Plane):
1) Administrator uses Admin UI → Management REST API → Config DB
2) DevOps pipelines use Helm/Docker CLI → push configuration via Management API
• Runtime flow (Data Plane):
1) Client request → Ingress/Nginx → Gateway runtime nodes
2) Policy Engine consults cache (Terracotta) and DB → enforces rate-limit, security, transformations
3) Gateway forwards request → target backend service → receives response
4) Transaction logs & metrics → Elasticsearch → Kibana
Architectural Patterns & Principles
• Control-Plane / Data-Plane separation to decouple management from runtime
• Externalize state in highly available stores (SQL/NoSQL for config, Elasticsearch for analytics, Terracotta for cache)
• Containerized deployments with declarative manifests (Docker Compose, Kubernetes, Helm)
• Extensibility points: new policies (JSON files under apigatewayservices), custom destinations, custom scripts
• Secure-by-design: integrate with external identity provider, TLS termination at ingress
Final Checklist
[ ] Components grouped by Control Plane, Data Plane, Data Stores, External Systems
[ ] Labels for technologies: Elasticsearch, Kibana, Terracotta, Nginx, Helm, Docker, Kubernetes
[ ] Notes on protocol (HTTPS, REST) and authentication (APIKey, OAuth)