Lab 4: Load Profiles

Learning Objectives

By the end of this lab, you'll understand:

• ✅ What load profiles are and why they matter
• ✅ How to implement smoke tests (minimal load)
• ✅ How to implement ramp tests (gradual increase)
• ✅ How to implement step tests (staircase increases)
• ✅ How to implement spike tests (sudden jumps)
• ✅ How to analyze results for each profile

Real-World Scenario

Your product team asks: "How does our system behave under different loads?"

You need to answer: - Smoke test: Can the system handle 1 user? (sanity check) - Ramp test: What happens as load gradually increases to 100 users? - Step test: At what point does performance degrade? (50 users → 100 → 200?) - Spike test: Can we handle a sudden viral moment (100 → 10,000 users instantly)?

Each load profile reveals different bottlenecks.

Concept: Load Profiles

Smoke Test

.injectOpen(
    constantUsersPerSec(1).during(60)  // 1 user for 1 minute
)

Purpose: "Is the system running?"

Ramp Test

.injectOpen(
    rampUsersPerSec(0).to(100).during(300)  // 0→100 users over 5 minutes
)

Purpose: "Where does it start struggling?"

Step Test

.injectOpen(
    constantUsersPerSec(50).during(60),
    rampUsersPerSec(50).to(100).during(60),
    constantUsersPerSec(100).during(60),
    rampUsersPerSec(100).to(150).during(60)
)

Purpose: "At which step does latency jump?"

Spike Test

.injectOpen(
    constantUsersPerSec(10).during(60),
    rampUsersPerSec(10).to(500).during(30),
    constantUsersPerSec(500).during(60),
    rampUsersPerSec(500).to(10).during(30)
)

Purpose: "Can we recover from sudden load?"

Running the Lab

cd /Users/sgovinda/Learn/GatlingLearning

# Smoke Test
mvn gatling:test -Dgatling.simulationClass=io.learn.gatling.simulations.http.Sim04_LoadProfileSmoke

# Ramp Test (5+ minutes)
mvn gatling:test -Dgatling.simulationClass=io.learn.gatling.simulations.http.Sim04_LoadProfileRamp

# Step Test (4+ minutes)
mvn gatling:test -Dgatling.simulationClass=io.learn.gatling.simulations.http.Sim04_LoadProfileStep

# Spike Test (3+ minutes)
mvn gatling:test -Dgatling.simulationClass=io.learn.gatling.simulations.http.Sim04_LoadProfileSpike

Expected Results

Key Takeaways

1. Smoke test = sanity check
2. Ramp test = performance curve
3. Step test = threshold identification
4. Spike test = resilience testing
5. Each profile reveals different insights

Navigation

← Previous: Lab 3: Checks & Validation
→ Next: Lab 5: CRUD Operations
↑ Up: Lab Overview