⚡ K6 Performance Testing Suite

Load & performance tests for the Daily Finance and DemoQA BookStore APIs

Assignment · Batch 18 · Topic: Performance Testing

---

📋 Table of Contents

• 🎯 Overview
• 🗂️ Project Structure
• ⚙️ Prerequisites & Setup
• 🚀 Quick Start
• 📘 Task 1 — Daily Finance
• 📗 Task 2 — DemoQA BookStore
• 📊 Results at a Glance
• 🎥 Submission Artifacts
• 🧠 Key Design Notes
• 🛠️ Troubleshooting

---

🎯 Overview

Two independent k6 test scripts, each producing a self-contained HTML report on completion:

#	Scenario	Focus	Load model
1	Daily Finance — admin authenticates, lists users, drills into one by ID	Functional correctness under load + per-endpoint thresholds	Constant 5 VUs · 30s
2	DemoQA BookStore — read the books catalogue	Response-time SLA under rising/peak/falling traffic	Staged ramp 0→10→25→0 VUs

Both scripts share the same reporting pipeline: the k6-reporter library renders a rich HTML dashboard via handleSummary(), while a live console summary is printed on stdout.

---

🗂️ Project Structure

K6Mastery/
├── task1-dailyfinance.js       # Task 1 — login → users → user-by-id (tagged + thresholds)
├── task2-bookstore.js          # Task 2 — ramping load against Books API
├── Reports/                    # 📄 Auto-generated HTML reports  (git-ignored)
│   ├── task1-dailyfinance-report.html
│   └── task2-bookstore-report.html
├── Screenshots/                # 🖼️ Report screenshots for submission
│   ├── task1-dailyfinance-report.png
│   └── task2-bookstore-report.png
├── .gitignore                  # Reports/ is excluded from version control
└── README.md

🔒 Reports/ is git-ignored as required by the assignment — reports are build artifacts, regenerated on every run.

---

⚙️ Prerequisites & Setup

1. Install k6 (the only hard requirement — scripts pull the HTML reporter from a CDN at runtime, so you just need internet access):

# macOS
brew install k6

# Windows
choco install k6

# Linux (Debian/Ubuntu)
sudo gpg -k && sudo apt-get install k6

2. Verify the install:

k6 version        # → k6 v2.1.0 (or newer)

That's it — no npm install, no local dependencies. ✅

---

🚀 Quick Start

# 📘 Task 1 — Daily Finance  (~30s)
k6 run task1-dailyfinance.js

# 📗 Task 2 — BookStore ramping load  (~2 min)
k6 run task2-bookstore.js

Each run writes its report to Reports/…-report.html. Open it in any browser:

open Reports/task1-dailyfinance-report.html    # macOS

💡 First run tip: the Reports/ folder must exist before k6 can write to it. If you cloned this fresh, run mkdir -p Reports once.

---

📘 Task 1 — Daily Finance (Authenticated User Lookup)

An admin logs in, pulls the full user list, randomly picks a user, re-fetches that user by ID, and asserts the round-trip is consistent.

🔁 Test Flow

sequenceDiagram
    autonumber
    participant K6 as k6 · Admin VU
    participant Auth as POST /auth/login
    participant List as GET /user/users
    participant One as GET /user/{id}

    K6->>Auth: { email, password }
    Auth-->>K6: 200 · { token }
    Note over K6: extract JWT token
    K6->>List: Authorization: Bearer «token»
    List-->>K6: 200 · [ 900 users ]
    Note over K6: pick a RANDOM user
    K6->>One: Bearer «token» · /user/«_id»
    One-->>K6: 200 · { user }
    Note over K6: ✔ assert _id matches<br/>🖨 print id · name · email · phone

Loading

🌐 APIs & Tags

Step	Method & Endpoint	k6 Tag
Login	POST /api/auth/login	{ name: "Admin Login API" }
List users	GET /api/user/users	{ name: "Get All Users API" }
User by ID	GET /api/user/{id}	{ name: "Get User By ID API" }

Base URL: https://dailyfinanceapi.roadtocareer.net

✅ Validations

• Login status is 200 and a token is present
• User list status is 200 and the list is a non-empty array
• User-by-ID status is 200
• 🎯 Selected user's id matches the searched user's id (the core integrity check)

📏 Per-Tag Thresholds — and how they held up

Metric (by tag)	Threshold	Measured p(95)	Result
http_req_duration{name:Admin Login API}	p(95) < 600ms	190 ms	🟢 Pass
http_req_duration{name:Get All Users API}	p(95) < 800ms	477 ms	🟢 Pass
http_req_duration{name:Get User By ID API}	p(95) < 600ms	146 ms	🟢 Pass
checks (overall success rate)	rate > 0.99	100%	🟢 Pass

🖨️ Sample console output (click to expand)

[Admin Login API]     status=200  response time=112.86 ms
[Get All Users API]   status=200  response time=277.94 ms  (users=900)
[Get User By ID API]  status=200  response time=65.14 ms
  --> Selected User:
        id          : 2878f6fb-f31a-4886-9379-c5da46bc349c
        name        : Jovan Romaguera
        email       : Sylvia899@gmail.com
        phoneNumber : 01575753529

     ✓ Login status is 200
     ✓ Token is present
     ✓ Get All Users status is 200
     ✓ User list is a non-empty array
     ✓ Get User By ID status is 200
     ✓ Selected user id matches searched user id

📸 HTML Report

309 requests · 0 failed · 0 breached thresholds · 0 failed checks

---

📗 Task 2 — DemoQA BookStore (Ramping Load)

Simulates traffic gradually increasing → holding at peak → gradually decreasing, verifying the Books API stays responsive throughout.

📈 Load Profile

xychart-beta
    title "Ramping VU Load Profile (120s total)"
    x-axis "Time (seconds)" [0, 20, 50, 70, 100, 120]
    y-axis "Virtual Users" 0 --> 25
    line [0, 10, 10, 25, 25, 0]

Loading

Stage	Duration	Target VUs	Behaviour
1	20s	10	🔼 Ramp up
2	30s	10	➡️ Hold
3	20s	25	🔼 Ramp up to peak
4	30s	25	➡️ Hold at peak
5	20s	0	🔽 Ramp down

Endpoint: GET https://demoqa.com/BookStore/v1/Books

✅ Validations

• Response status is 200
• Response body contains books (non-empty array)
• Response time is < 600ms (per-request check)
• 🖨️ Prints the total number of books returned (surfaced both to console and as a custom books_returned gauge in the report)

📏 Thresholds — and how they held up

Metric	Threshold	Measured	Result
http_req_duration	p(95) < 600ms	563 ms	🟢 Pass
http_req_failed	rate < 1%	0.00%	🟢 Pass

📚 Total books returned: 8. Under peak load a handful of individual requests crossed the 600 ms mark (per-request checks landed at 99.7%, max 1.12s), yet the aggregate p(95) stayed under 600 ms — exactly the SLA the test is designed to guard.

🖨️ Sample console output (click to expand)

Total number of books returned from the API: 8

     ✓ status is 200
     ✓ body contains books
     ✗ response time < 600ms  ↳ 99% — ✓ 1397 / ✗ 12

   ✓ http_req_duration...: p(95)=563.28ms   (threshold p95<600 ✔)
   ✓ http_req_failed.....: 0.00%            (threshold rate<0.01 ✔)
     books_returned......: 8   min=8  max=8

📸 HTML Report

1,409 requests · 0 failed requests · 0 breached thresholds

---

📊 Results at a Glance

	Task 1 · Daily Finance	Task 2 · BookStore
Requests	309	1,409
Failed requests	0 (0.00%)	0 (0.00%)
Breached thresholds	0 ✅	0 ✅
Checks passed	100%	99.7%
Peak load	5 VUs	25 VUs
Slowest p95	477 ms	563 ms
Verdict	🟢 PASS	🟢 PASS

---

🎥 Submission Artifacts

Artifact	Location	Status
Task 1 HTML report	Reports/task1-dailyfinance-report.html	✅
Task 2 HTML report	Reports/task2-bookstore-report.html	✅
Task 1 report screenshot	Screenshots/task1-dailyfinance-report.png	✅
Task 2 report screenshot	Screenshots/task2-bookstore-report.png	✅
🎬 Screen recording of both runs	record locally	⬜

To record the run video (required for submission): start a screen recorder — on macOS, QuickTime → File → New Screen Recording — then run each command and let the report open:

k6 run task1-dailyfinance.js
k6 run task2-bookstore.js

---

🧠 Key Design Notes

These are the deliberate decisions worth knowing before grading.

• name & id field mapping. The Daily Finance API returns _id, firstName, and lastName — there is no single name or id field. The script therefore maps id → _id and prints name as firstName + lastName, matching the assignment's intent.

• Tag names aligned to threshold selectors. k6 per-tag thresholds only receive data when the tag string exactly matches the threshold key. The assignment's threshold block uses the … API suffix (e.g. name:Admin Login API), so the request tags use the identical names. This guarantees the thresholds actually evaluate real samples instead of silently passing on zero data.

• HTML report generation. handleSummary() returns the report to Reports/…-report.html via k6-reporter, plus a colourised console summary via k6-summary — no external tooling or post-processing needed.

• Book count surfaced twice. Task 2 logs the total book count to the console and records it as a custom books_returned gauge, so the number is visible in the HTML report as well as stdout.

---

🛠️ Troubleshooting

Symptom	Cause	Fix
could not open 'Reports/…': no such file or directory	Reports/ folder doesn't exist yet	mkdir -p Reports before running
HTML report import fails / hangs	No internet access	The reporter is fetched from a CDN — connect to the internet
Task 1 thresholds show no data	Tag names edited out of sync with threshold keys	Keep request tags identical to the threshold selectors
k6: command not found	k6 not installed / not on PATH	See Prerequisites & Setup

---

Batch 18 · Performance Testing Assignment Built with k6 ⚡ · Reports by k6-reporter