LoadRunner: Controller, Scenarios & Analysis

The Controller & Load Generators

The Controller orchestrates a test scenario -- how many VUsers run which scripts, and how they're distributed across Load Generators (the machines that actually execute VUsers, since one machine often can't generate enough load alone).

Ramp-Up

Rather than starting all VUsers instantly, ramp-up gradually increases active VUsers over a defined period -- mimicking how real traffic builds up, and helping pinpoint the load level where problems first appear.

Test Scenario Types

• Load testing -- expected peak concurrent usage, verifying acceptable performance.

• Stress testing -- pushing beyond expected peak to find the actual breaking point.

• Soak/endurance testing -- sustained moderate load over hours or days, surfacing issues like memory leaks that only appear over time.

Transactions: Measuring What Matters

A Transaction wraps a meaningful business step ("Login", "Add to Cart") with start/end markers, so Analysis reports response time for that specific step -- not just an undifferentiated whole-script time.

Analysis & Baselines

After a run, Analysis turns collected metrics into graphs and reports for investigating bottlenecks -- correlating application response times with infrastructure monitoring (CPU, DB query time) often reveals which tier is the actual constraint. Establishing a baseline run under known conditions gives a reference point for whether later changes actually helped or hurt.

Testing Against Staging, Not Production

Teams typically run heavy load tests against a production-like staging environment rather than production itself, to get representative data without risking degrading the experience for real users.