Ship faster. Block deploys when load tests fail.
LoadGen runs as the load-test gate inside your CI/CD pipeline — visual API flows, GitOps sync, and a typed REST API surface that any pipeline runner can call.
Pipeline Run · #4218
Triggered by Jenkins via POST /api/testing/runs.
{
"profileId": "prod-gateway-test",
"scenarioId": "stress-test-v1",
"variables": {
"target": "staging.lg.com"
}
}The Problem
Pipelines without load gates are pipelines that ship regressions.
Continuous deployment is fast. Continuous deployment without a performance gate is fast and reckless.
Manual gates slow your shipping cadence.
A release that has to wait for a human to read a Grafana dashboard isn’t a continuous deployment — it’s a queue with extra steps.
Script-based load tests rot in CI configs.
jmx files in `.github/workflows/` are nobody’s idea of source-of-truth — they drift the moment the platform team rotates.
Capacity drift only surfaces in production.
Without a load-test gate in the pipeline, every regression is shipped, and the only signal that something is wrong is the on-call page at 02:00.
How LoadGen solves this
Define · Sync · Run · Gate.
The same engine that authors flows in the UI runs them headless in CI — same JSON, same auth, same results.
Define Flow
Author API and load flows visually — no scripts. OpenAPI / Swagger import, environment variables, reusable auth profiles.
Sync to Git
Push collections and workloads to the central Git repository. Version every change. Diff before merge. Roll back as a commit.
Run in Pipeline
Trigger from Jenkins / Azure DevOps / GitHub Actions / GitLab CI via the typed REST API. JWT auth. Server-side BFF pattern.
Gate the Deploy
Block promotion when p95, error-rate, or capacity thresholds breach. Pass through cleanly when the run is green.
Live · Pipeline Run
The gate stage your pipeline was missing.
Build, lint, test, integrate, hit the API surface, run the load gate, and only then promote. Every stage emits the same JSON; every stage is callable from any runner.
Outcomes · before / after
Faster deploys. Fewer regressions. Quieter on-call.
Drawn from teams that adopted the load-test gate as a required pipeline stage for one quarter.
Before
4 h 12 m
After
22 min
Before
14 %
After
1.2 %
Before
8
After
1
Before
28 h
After
4 h
Pipeline → Platform
Four features. One pipeline gate.
CI/CD integration isn’t a single feature — it’s the way these four compose into a passing-or-failing gate stage.
- /api-testing/flows— Visual flow editor
- /api-testing/runs— Pipeline run history
- /sync/overview— GitOps console
- /testing/schedules— Cron baselines
- /api-testing/environments— Per-stage config
API Flows
Drag-and-drop API validation with 28+ specialized node types. OpenAPI import, env-aware auth.
/features/flow-editor
Sync
GitOps for collections and workloads. Push, pull, diff, history — same model as your application code.
/features/sync
Agents
Right-sized agents per workload — Core for web, Full for desktop, VDI for AVD / Horizon.
/features/agents
Alerting
When the gate fails, the right channel pages — email, webhook, SMS, WhatsApp.
/features/alerting
Make the load test the gate, not the post-mortem.
Wire LoadGen into your pipeline this week. Block the next regression. Ship the release after with the same green checkmark every other stage gives you.
