Enterprise Service Virtualization & High-Throughput HTTP Mocking Harness

Senior Automation Engineer · Year 2 · Q6 (Previous SaaS)

Platform & InfrastructureYear 2 · Q6 (Previous SaaS)

Architected an enterprise service virtualization engine to mock third-party payment gateways and legacy SOAP/REST dependencies, cutting sandbox vendor fees by $85,000 annually.

⚡12,000+ Mock Requests/Sec Throughput
⚡$85,000 Annual Vendor Sandbox Cost Saved
⚡Zero Test Failures from Rate Limiting
JavaWireMockNettyRedisDockerJackson

Situation

Integration test suites depended on third-party staging sandboxes (e.g., payment gateways, SMS providers, credit checking services). These external endpoints introduced severe rate-limiting, frequent weekend maintenance downtime, and per-API-call billing costs that scaled linearly with pipeline parallelization.

Task

Design a high-throughput, stateful HTTP service virtualization harness capable of simulating third-party network failures, latency spikes, and dynamic payload responses locally inside Docker containerized build pipelines.

Action

  • Non-Blocking I/O Architecture: Wrapped a headless WireMock core in a custom Netty-based server wrapper, optimizing socket reuse and thread pools to process over 12,000 HTTP requests per second.
  • Stateful Response Stubs: Engineered a dynamic scenario engine backed by Redis that maintained stateful transaction workflows (e.g., Authorizing -> Captured -> Refunded) across multi-step API sequences without external network calls.
  • Fault Injection Profiles: Implemented dynamic HTTP header injection (X-Simulate-Fault: connection-drop) to force the mock engine to drop connections mid-transfer or inject 5,000ms latency to validate client-side timeout logic.

Result

  • Saved $85,000 per year in third-party API sandbox execution costs.
  • Completely eliminated third-party API rate-limiting errors in CI, reducing test execution time by 35%.