Checkpoint Lab — HTTP Request Samplers, Defaults, Headers, Cookies, and Cache Managers

1. Assumptions and hard safety limits

2. Prepare the workspace and fixture

jmeter-ch05-checkpoint/
├── fixtures/
│   └── http_fixture.py
├── plans/
│   ├── checkpoint-debug.jmx
│   └── checkpoint-load.jmx
├── tools/
│   └── analyze_jtl.py
├── evidence/
│   ├── run-manifest.txt
│   ├── predictions.txt
│   └── validity.txt
└── results/
    └── checkpoint-001/

Start a fresh fixture with a fresh event log:

python fixtures/http_fixture.py --log results/checkpoint-001/server-events.jsonl
curl --fail --silent http://127.0.0.1:8000/health
curl --fail --silent http://127.0.0.1:8000/stats

3. Build the checkpoint tree

Test Plan — Chapter 05 Checkpoint
└── Thread Group — 2 users × 1 loop
    ├── HTTP Request Defaults
    │   ├── Implementation: HttpClient4
    │   ├── Protocol: http
    │   ├── Server: 127.0.0.1
    │   ├── Port: 8000
    │   ├── Connect timeout: 1000 ms
    │   └── Response timeout: 2000 ms
    ├── HTTP Header Manager
    │   ├── Accept: application/json
    │   └── X-Lab-Client: devops-academy-ch05
    ├── HTTP Cookie Manager — standard; no manual cookies
    ├── HTTP Cache Manager — Cache-Control/Expires ON; max 100
    ├── POST Session — /session
    │   └── HTTP Header Manager — Content-Type: application/json
    ├── GET Profile — /profile
    ├── GET Cache First — /cache
    ├── GET Cache Second — /cache
    ├── GET Redirect — /redirect; Follow Redirects ON
    └── View Results Tree — DEBUG COPY ONLY

Use KeepAlive on all HTTP samplers. Keep Redirect Automatically off.

4. Write predictions before execution

1. Two POST Session requests create two distinct synthetic session cookies.
2. Each Profile request should carry the cookie stored by that thread and return HTTP 200.
3. Each thread has its own Cache Manager state. The first Cache GET per user should populate that user's cache; the second may be satisfied from that user's fresh cache without a second target hit.
4. Each Redirect sampler should cause a /redirect request followed by a /final request because Follow Redirects is enabled.
5. Client ports may show connection reuse within a user's sequence, but exact reuse is an observation rather than a required pass/fail condition.

5. Bounded debug verification

Run checkpoint-debug.jmx once in GUI mode only. In View Results Tree:

• inspect Session response Set-Cookie and synthetic session JSON;
• verify Profile is HTTP 200 rather than 401;
• inspect Cache First/Second behavior and server counters;
• inspect Redirect parent/children;
• record request/response metadata needed to explain the run, then close/disable the listener.

Do not retain screenshots containing real credentials—there are none in this fixture by design.

6. Prove per-thread state from server evidence

Inspect server-events.jsonl. You should see two session-creation POSTs and two Profile GETs. Profile events should contain session cookies issued by earlier Session responses. The exact request interleaving can differ because two threads run concurrently.

Query /stats and record sessions_created=2. Cache path counts may be lower than the number of Cache sampler executions because fresh Cache Manager entries can avoid target calls.

7. Create the lean load-evidence copy

Save checkpoint-load.jmx with View Results Tree disabled. Do not add Debug Sampler or response-saving listeners. Restart the fixture/event log so evidence has a clean baseline.

8. Run from CLI

jmeter -n   -t plans/checkpoint-load.jmx   -l results/checkpoint-001/results.jtl   -j results/checkpoint-001/jmeter.log   -Jjmeter.save.saveservice.print_field_names=true   -Jjmeter.save.saveservice.connect_time=true   -Jjmeter.save.saveservice.response_data=false   -Jjmeter.save.saveservice.response_data.on_error=false   -Jjmeter.save.saveservice.requestHeaders=false   -Jjmeter.save.saveservice.responseHeaders=false   -Jjmeter.save.saveservice.samplerData=false
python tools/analyze_jtl.py results/checkpoint-001/results.jtl

PowerShell:

jmeter.bat -n `
  -t plans\checkpoint-load.jmx `
  -l results\checkpoint-001\results.jtl `
  -j results\checkpoint-001\jmeter.log `
  -Jjmeter.save.saveservice.print_field_names=true `
  -Jjmeter.save.saveservice.connect_time=true `
  -Jjmeter.save.saveservice.response_data=false `
  -Jjmeter.save.saveservice.response_data.on_error=false `
  -Jjmeter.save.saveservice.requestHeaders=false `
  -Jjmeter.save.saveservice.responseHeaders=false `
  -Jjmeter.save.saveservice.samplerData=false
python tools\analyze_jtl.py results\checkpoint-001\results.jtl

9. Verification checklist

• JTL contains the expected sampler labels and no unexpected failures.
• Profile samples return HTTP 200 for both users.
• Server event log proves two independent server-issued session cookies were later presented.
• Server cache path count is interpreted in light of per-thread Cache Manager state, not assumed from sampler count alone.
• Redirect target log shows both /redirect and /final.
• jmeter.log contains no unexpected engine/configuration errors.
• JTL does not contain response bodies/request headers/response headers.
• Generator CPU/memory/network observation shows this tiny run had safe headroom.

10. Optional one-thread embedded-resource evidence

Run a separate one-thread plan with GET Page first with embedded resources off, then on. With it on and concurrent pool off, the server event log should add requests for /asset.css, /asset.js, and /pixel.svg.

Record explicitly that the script file was downloaded but not executed. This avoids contaminating the main cookie/cache checkpoint with a different request population.

11. Run manifest

run_id=chapter05-checkpoint-001
jmeter=5.6.3
java=17
http_implementation=HttpClient4
target=http://127.0.0.1:8000
threads=2
loops=1
cookie_manager=standard; no manual cookies
cache_manager=Cache-Control/Expires enabled; max=100
redirect=Follow Redirects ON; Redirect Automatically OFF
keep_alive=ON
embedded_resources=OFF in main checkpoint
jtl=lean CSV; bodies/headers disabled; connect_time enabled
abort=target mismatch, external URL, unexpected 5xx, unsafe generator pressure

12. Required evidence packet

13. Validity statement

14. Cleanup and rollback

1. Stop the local Python fixture.
2. Keep the checkpoint artifacts until review is complete.
3. Remove/disable View Results Tree before carrying the JMX forward.
4. Do not carry synthetic cookies as manually defined Cookie Manager entries.
5. No production target, real credential, certificate, database, remote engine, plugin, container, or system-wide network/JVM setting was changed.

15. What Chapter 05 adds to the operating model

Chapter 04 defined when users act. Chapter 05 adds a protocol/session contract: resolved destination, HTTP client implementation, request headers, server-issued cookie lifecycle, per-user cache behavior, redirect policy, connection behavior, embedded-resource scope, and result-retention policy.

Chapter 06 builds on this by separating JMeter variables, JMeter properties, Java system properties, and environment-specific configuration so the same HTTP journey can move safely between authorized environments without hard-coded values or shared mutable state.

Knowledge check

Official references and version notes

• Component Reference — current HTTP Request, HTTP Request Defaults, HTTP Header Manager, HTTP Cookie Manager, HTTP Cache Manager, redirect, keep-alive, and embedded-resource semantics.
• Elements of a Test Plan — scope and execution architecture around samplers and configuration elements.
• Properties Reference — current CSV/JTL save fields and defaults such as response/request header retention and connect time.
• Functions and Variables — thread-local runtime values and function boundaries.
• Getting Started — GUI authoring/debugging versus CLI load execution.
• Best Practices — lean load generation and result/listener guidance.
• Apache JMeter downloads — current production release and release-specific Java requirement.