Terminologies · Performance Testing 101

Performance Testing Terminologies

The essential vocabulary every performance tester must know — from Virtual Users and Response Time to Throughput, Pacing, and how they all connect.

Virtual User (VU)

👤

A simulated user generated by a performance testing tool. It behaves exactly like a real user — sending requests, waiting for responses, entering data, navigating pages.

Testing 500 users logging into an e-commerce site → tool creates 500 VUs, each performing: Open homepage → Enter credentials → Click login → Add product.

⚙️ JMeter: Each Thread = 1 Virtual User

Response Time

⏱️

Total time from the moment the request leaves the client until the last byte of the response is received. Includes network transfer, server processing, and intermediate component delays.

Network Transfer Server Processing Load Balancer DB / Cache API Gateway

Response Time = Time(Last Byte Received) − Time(First Byte Sent)

Request sent at 10:00:01.000 · Last byte received at 10:00:02.200 → Response Time = 1.2 seconds

Throughput

📊

Number of requests or transactions the system successfully processes per second/minute. It measures the capacity of the system. Higher throughput = system handles more load.

Requests/sec (RPS) Transactions/sec (TPS) Pages/min KB/s · MB/s

During peak: API handles 150 req/sec · Checkout handles 25 TPS

Request

📤

A single call sent from client to server. The atomic unit of communication.

⚙️ JMeter: Each HTTP Request Sampler = 1 Request

Transaction

🔗

A group of related requests that represent one business action. Used to measure response time of the entire workflow.

"Login" Transaction = 3 requests:
① Load login page → ② Submit credentials → ③ Load dashboard
→ Gives you the overall response time of the Login workflow.

⚙️ JMeter: Use Transaction Controller to group requests

Iteration

🔄

One complete end-to-end journey of the user. Every Virtual User repeatedly executes the script scenario — each run is an iteration.

E-commerce flow: Login → Search → View Product → Add to Cart → Checkout → Logout = 1 Iteration.
1 hour test, user repeats 20 times → 20 iterations.

Think Time

💭

The pause a user takes between actions. Simulates real human behavior. Without think time, users are unrealistically fast → inaccurate load test results.

Real user reads product details → waits 8s · Chooses size → waits 4s · Reviews cart → waits 6s

⚙️ JMeter: Use Constant / Gaussian / Uniform Timers

Pacing

⏲️

The delay between two complete iterations. Controls how frequently each user repeats the full scenario. Used to achieve the required TPS/RPS.

Users = Iterations/sec × Pacing (Little's Law)

Each user performs 1 full iteration every 30 seconds → Pacing = 30s. Without pacing, TPS becomes too high or unrealistic.

🎯

Putting It All Together

Scenario

Test an e-commerce site with 200 virtual users. Each user performs a full business flow with 5s think time between pages and 20s pacing.

1 Iteration = Full User Journey

👤 Login

→

🔍 Search

→

🛒 Add to Cart

→

💳 Checkout

200 Virtual Users

5s Think Time

20s Pacing

80 TPS Throughput

1.5s Avg Response Time