Table of Contents
Table of Contents
Further Reading
No items found.
No items found.
No items found.
No items found.
No items found.
No items found.
What Is Decoupled Testing?
Decoupled testing separates test execution from CI/CD pipelines. Build tools like Jenkins or GitHub Actions no longer manage every test.
Instead, an independent orchestration layer runs tests directly in your infrastructure. That improves flexibility, scale, and reliability.
Decoupled vs. CI-Integrated Testing
Why Decoupled Testing Matters
CI/CD pipelines become bottlenecks as suites grow. Tying tests to builds adds complexity, slows deployments, and limits scale.
- Flexible execution: run on demand, on schedules, or via API
- Independent scale: tests grow without adding build capacity
- Reusable configs: share definitions across environments
- Faster, steadier pipelines: less load and fewer timeouts
Testing then runs alongside development instead of as a final pipeline stage.
How Decoupled Testing Works
- Define separately: version tests apart from pipeline files.
- Trigger anywhere: start runs from APIs, webhooks, schedules, or CI steps.
- Execute in Kubernetes: orchestration scales across clusters and environments.
- Centralize results: logs and outcomes land in one dashboard for visibility and governance.
Real-World Examples
- B2B SaaS microservices: teams run contract and integration tests in Kubernetes instead of CI pipelines
- Enterprise QA: security and performance tests run in isolated environments or off-peak
- Developer testing: engineers trigger specific suites from the CLI without full builds
How It Relates to Testkube
Testkube is built around decoupled testing. It:
- Runs natively in Kubernetes: independent of any CI/CD tool
- Offers flexible triggers: CLI, API, dashboard, schedules, and events
- Uses versioned definitions: Git-based workflows for test configuration
- Reports back to CI: pipelines can still gate on results
Teams keep CI/CD-agnostic testing while achieving continuous coverage.
Best Practices
- Version tests: keep definitions in Git and environment-agnostic
- Use API or CLI triggers: repeatable, scriptable execution
- Centralize dashboards: visibility across clusters
- Separate concerns: keep orchestration apart from build and deploy logic
Common Pitfalls
- Over-coupling: tests tied to specific pipeline stages
- Poor visibility: decoupled tools without central management
- Inconsistent data: test data that differs across environments