LOLC Technologies · FusionX
FusionX: Automation, Performance & Containerised Test Envs
End-to-end web automation, JMeter-driven performance testing, and Docker + Kubernetes-based test environments for FusionX, with Grafana dashboards translating load-test signals into actionable engineering insight.
Problem
FusionX had to hold up under real load, not just pass its unit tests. Functional coverage was one axis; performance was another; and both suffered from environment drift: tests behaved differently depending on whose laptop ran them. We needed regression, load, and environment reproducibility to move together.
Architecture
Three connected layers, one operational story:
- Functional web automation: Selenium + Java suites covering the critical user journeys and core business workflows, authored against the same POM discipline used across LOLC's QA org.
- Performance & load: JMeter test plans simulating high transaction volume against FusionX, evaluating throughput, response time, and system reliability under stress. Grafana dashboards translate raw JTL streams into signals the engineering team can act on.
- Containerised environments: Docker images for regression and performance runners, orchestrated via Kubernetes. The same suite runs the same way on a developer's machine, in CI, and in a scaled-up load rig.
My contributions
- Developed Selenium + Java automation frameworks covering critical user journeys and core business workflows for FusionX.
- Authored full JMeter performance and load test plans; analysed bottlenecks; produced Grafana reporting dashboards that translated results into engineering action.
- Containerised regression and performance test environments with Docker + Kubernetes, improving scalability, consistency, and execution efficiency across CI pipelines.
- Managed end-to-end manual test execution cycles in Xray with defect lifecycle tracked in JIRA, so functional and non-functional coverage stayed traceable from requirement to release milestone.
Outcome
FusionX gained continuous evidence across functional, performance, and environmental layers that survived the transition from a laptop-run smoke test to a production-scale release process. Performance regressions surfaced with a graph, not a hunch. And test environments stopped being a distinct troubleshooting category all their own.
TODO: add p95 latency numbers and load-run screenshots from Grafana once approved for external sharing.