⚙️ Basic of JMeter

Created By: Shadikun Nahar Sakiba

1. What is JMeter?
●​ Apache JMeter is an open-source tool used for performance testing and
load testing of web applications, APIs, databases, and other services.​

●​ It simulates multiple virtual users sending requests to your server to

measure performance, throughput, and reliability.​

●​ Can test:​

○​ Web applications (HTTP/HTTPS)​

○​ SOAP/REST APIs​

○​ Databases (via JDBC)​

○​ FTP, JMS, Mail, LDAP​

○​ TCP services​

2. Key JMeter Components

a. Test Plan

●​ The main container in JMeter.​

●​ Holds all threads, samplers, listeners, and configurations.​

●​ You can parameterize and control execution flow here.​

b. Thread Group

●​ Represents a group of virtual users.​

●​ Each thread is one virtual user.​

●​ Key parameters:​

1.​ Number of Threads (Users): Number of virtual users to simulate.​

2.​ Ramp-Up Period (seconds): Time to start all users gradually.​

■​ Formula: Ramp-Up / Threads = Time between

starting each user​

■​ Example: 10 users, 20 sec ramp-up → 1 user starts every 2

seconds.​

3.​ Loop Count: Number of times each thread executes the test.​

4.​ Scheduler: Can schedule test duration.​

c. Sampler

●​ Defines what action the thread will perform.​

●​ Examples:​

○​ HTTP Request → Calls a URL.​

○​ JDBC Request → Executes a SQL query.​

○​ FTP Request → Upload/download files.​

d. Listener

●​ Used to view and analyze test results.​

●​ Examples:​

○​ View Results Tree: Shows request/response data.​

○​ Summary Report: Shows aggregated stats.​

○​ Graph Results / Response Time Graph: Visualizes trends.​

e. Config Elements

●​ Provide data or configurations to samplers.​

●​ Examples:​

○​ HTTP Request Defaults: Set default server/path.​

○​ CSV Data Set Config: Parameterize requests with external data

(usernames, IDs).​

f. Timer

●​ Introduces delay between requests.​

●​ Helps simulate real-world user behavior.​

g. Assertion

●​ Used to validate responses.​

●​ Examples:​

○​ Response contains a string.​

○​ Response code is 200.​

h. PreProcessor / PostProcessor

●​ PreProcessor: Runs before request (e.g., add headers, encode data).

A Pre-Processor in JMeter is a component that runs before an HTTP

Request (or sampler) is sent.​
Its purpose is to manipulate data, variables, headers, or the request itself
dynamically for each iteration or thread.

✅ Key Use Cases:

●​ Add dynamic values (random numbers, timestamps) to requests.​

●​ Modify headers or body before sending.​

●​ Prepare variables for the upcoming request.

●​ PostProcessor: Runs after request (e.g., extract token using regex).

A Post-Processor is a JMeter component that runs after an HTTP request

(sampler) is executed.​
Its main purpose is to extract or process data from the server’s response
so it can be used in subsequent requests.
 ✅ Key Use Cases:
●​ Extract IDs, tokens, session values from responses.​

●​ Validate dynamic response content.​

●​ Feed data into the next HTTP requests in the workflow.

3. Important JMeter Terms

a. Thread

●​ A virtual user in JMeter.​

●​ Executes samplers independently.​

●​ Multiple threads = concurrent users.​

b. Throughput

●​ Number of requests processed per unit time.​

●​ Unit: requests/sec or requests/minute.​

●​ High throughput = server can handle many users efficiently.​

c. Ramp-Up

●​ Time to start all threads.​

●​ Helps avoid sudden load spike.​

 ●​ Formula example:​
Ramp-Up = 10 sec, Threads = 5 → 1 thread starts every
2 sec​

d. Response Time

●​ Time taken by server to respond to a request.​

●​ Lower is better.​

●​ Types:​

○​ Average Response Time​

○​ Median Response Time​

○​ 90th Percentile (time within which 90% of requests complete)​

e. Sample Time

●​ Time taken to complete one sampler request.​

●​ Measured for each thread separately.​

f. Latency

●​ Time from request sent to first byte received.​

●​ Useful to check server/network delays.​

**g. Error %

●​ Percentage of requests that failed or returned errors.​

 ●​ Formula:​
Error % = (Number of failed requests / Total requests)
* 100​

h. Threshold

●​ Performance limit set for response times, errors, or throughput.​

●​ Example: “All requests must respond within 2 sec” → Threshold = 2 sec.​

i. Load / Stress Testing

●​ Load Testing: Test application at expected user load.​

●​ Stress Testing: Test application beyond max capacity to see failure points.​

4. Metrics JMeter Shows

Metric Meaning

Throughput Requests/sec handled by

server

Response Time server takes to respond

Time

Latency Time to first byte

Sample Number of requests sent
Count

Error Count / Number/percentage of failed

% requests

KB/sec Bandwidth used during test

5. JMeter Execution Modes

1.​ GUI Mode: For designing and debugging test plans.​

Non-GUI Mode: For actual load testing and report generation.​

​
jmeter -n -t [Link] -l [Link] -e -o
report-folder

6. JMeter Functions
●​ ${__Random(1,100)} → Generate random number.​

●​ ${__time()} → Current timestamp.​

●​ ${__UUID()} → Generate unique ID.​

●​ ${__CSVRead([Link],0)} → Read data from CSV file.​

 ●​ ${__BeanShell()} or ${__groovy()} → Custom scripts.​

Functions allow dynamic data and parameterization.

Ramp Up Time calculation:

🎯 Example Setup
Setting Value

Number of Threads (users) 10

Ramp-Up Period (seconds) 20

🧮 What JMeter Does Internally

JMeter must start 10 users within 20 seconds.

That means:

Ramp-Up Time per User=Ramp-Up PeriodNumber of Users=2010=2 seconds\text{Ramp-Up

Time per User} = \frac{\text{Ramp-Up Period}}{\text{Number of Users}} = \frac{20}{10} = 2 \text{
seconds}Ramp-Up Time per User=Number of UsersRamp-Up Period​=1020​=2 seconds

✅ So, every 2 seconds, JMeter will start one new virtual user.

🕒 Timeline Example
Time Action
(sec)

0 User 1 starts

2 User 2 starts

4 User 3 starts

6 User 4 starts

8 User 5 starts
 10 User 6 starts

12 User 7 starts

14 User 8 starts

16 User 9 starts

18 User 10 starts

20 All 10 users are now running concurrently

📘 What This Is Called

This calculation or concept is known as the “Ramp-Up Rate” or “User Start Interval.”

Formula:

Ramp-Up Rate=Ramp-Up PeriodNumber of Threads\text{Ramp-Up Rate} = \frac{\text{Ramp-Up

Period}}{\text{Number of Threads}}Ramp-Up Rate=Number of ThreadsRamp-Up Period​

It tells you how often (in seconds) JMeter will launch a new user.

🧩 Your Setup
Setting Value

Threads (Users) 10

Ramp-Up Period 20
seconds

Loop Count 2

🔹 1️⃣ Meaning of “Ramp-Up Period = 20 seconds”

This doesn’t mean the whole test will finish in 20 seconds.​
It means:

“It will take 20 seconds just to start all 10 users.”

So, by 20 seconds, all 10 virtual users are active (started),​
but they are still running their tests (executing HTTP requests).

🔹 2️⃣ How Ramp-Up Works in Time

Each user starts 2 seconds apart:

20 sec ramp-up÷10 users=2 sec per user20\text{ sec ramp-up} ÷ 10\text{ users} = 2\text{ sec
per user}20 sec ramp-up÷10 users=2 sec per user
Time User Action
(sec)

0 User 1 starts

2 User 2 starts

4 User 3 starts

... ... ...

18 User 10 starts

20 ✅ All users are now active

🔹 3️⃣ What Happens After Ramp-Up

Now all 10 users are running their loops.

Each user:

●​ runs Loop 1​

●​ then immediately runs Loop 2​

●​ independently, based on when they started​

So the total test time depends on:

1.​ How long your requests take (the duration of each loop)​
 2.​ Ramp-up time​

🧮 Let’s Take an Example

Let’s assume:

●​ Each loop (one full test) = 10 seconds​

Then:

User Start Time Loop 1 Loop 2 Ends At

(10s) (10s)

User 1 0s 0–10s 10–20s 20s

User 2 2s 2–12s 12–22s 22s

User 3 4s 4–14s 14–24s 24s

… … … … …

User 18s 18–28s 28–38s 38s

10

🧠 So, What’s Actually Happening?

●​ Ramp-Up = just the time to start all users (not total test time)​

●​ Loop 2 starts after Loop 1 finishes for each user​

●​ The last user starts late (at 18s), so the test will continue beyond 20 seconds​

⚙️ Total Test Duration ≈ (Last user start time + Last user

response time) - (First user start time)
(but could be a bit more because last user starts late)

1️⃣ What “Each Loop Duration” Means

●​ In JMeter, each loop = the time it takes for one user to complete all samplers
(requests) in the Thread Group once.​

●​ This includes:​

1.​ All HTTP requests or other samplers in that Thread Group​

2.​ Think time / timers (if any)​

3.​ Pre-processors / post-processors / assertions​

So “10 seconds per loop” is not a fixed number — it depends on your test plan
and server response times.

🔹 2️⃣ How to Find Loop Duration in JMeter

Method 1: Using “View Results in Table”

1.​ Add Listener → View Results in Table to Thread Group​

2.​ Run the test​

3.​ Check the “Time” column for each sampler​

4.​ Sum up all samplers in one iteration → gives approximate loop duration per user​

Method 2: Using “Summary Report”

1.​ Add Listener → Summary Report​

2.​ Run the test​

 3.​ The “Average” column shows average response time per request​

4.​ Multiply by number of requests per loop → gives loop duration​

Method 3: Use a Timer / Stopwatch

●​ You can manually note the start time of first request and end time of last request for
a loop​

●​ Subtract: Loop Duration = End Time – Start Time​

🔹 3️⃣ Practical Example

Thread Group has 3 HTTP requests:

Reques Avg Response

t Time

A 2 sec

B 1.5 sec

C 2.5 sec

●​ ​
Loop Duration ≈ 2 + 1.5 + 2.5 = 6 seconds​

●​ If Loop Count = 2 → user takes ≈ 12 sec for 2 loops​

If you have timers (think time) → add that to the duration