Skip to content

alwayspulkit/FlakeyTestDashboard

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

4 Commits
 
 
 
 

Repository files navigation

Flaky Test Dashboard

An AI-powered test failure analysis tool that helps QA teams identify root causes, track trends, and prioritize fixes across Android, iOS, and Web platforms.


Table of Contents


Overview

Capability Detail
Test capacity 2,000+ with smooth performance
AI confidence 60–95% for root cause detection
Platforms Android, iOS, Web
Trend tracking 7-day vs 7-day comparison
Grouping algorithms 5+ (error message, stack trace, name, error code, pattern)
Dependencies None — pure HTML/CSS/JavaScript

Quick Start

1. Open flaky-test-dashboard.html in any modern browser.

2. Load data — choose one:

  • Click 📊 Load Demo Data to generate 2,000 realistic test records instantly.
  • Click 📂 Upload CSV File and select your file (see CSV format below).

3. Explore — the dashboard auto-analyzes your data and populates all sections.

4. Act — use Create Ticket buttons to log issues, and Trend Analysis to verify fixes over time.


Core Features

Quick Stats

Four summary cards at the top of the page give you an instant snapshot:

Card What it shows
Total Tests Count of all failures in current filter
AI Groups Number of detected failure clusters
Top Platform Most affected platform with icon
Trend Direction and % change vs previous period

AI-Grouped Failures

Main section — left panel

Similar failures are automatically clustered using five algorithms:

  • Error Message Similarity — Levenshtein distance between error strings
  • Stack Trace Similarity — function call signature comparison
  • Test Name Similarity — normalized name matching
  • Error Code Clustering — HTTP status code grouping
  • Pattern Recognition — regex-based pattern matching

Each group card shows:

  • Pattern name, affected test count, and platform badges
  • Severity: 🔴 Critical (10+ tests) · 🟠 High (5–10) · 🟡 Medium (1–5)
  • Root cause analysis box when confidence exceeds 60%
  • Expandable list of individual tests
  • Create Ticket for Group button

Groups are paginated at 10 per page.


Smart Insights

Main section — right panel

AI-generated recommendations surfaced from your failure patterns:

Insight Type Example
Platform Alert "Android has 2.3x more failures than iOS"
Team Impact "Auth Team most affected — 450 failures"
Trend Warning "Failures increased 120% this week"
Quick Win "Fix 1 root cause to resolve 450 tests (95% confidence)"

Trend Analysis

Full-width section — bottom of page · 2-column layout

Compares the last 7 days against the prior 7 days per root cause.

Trend categories:

Direction Criteria Border Urgency
📈 Increasing ≥ +20% Red 🚨 Critical (≥+50%) or ⚠️ High (+20–50%)
📉 Decreasing ≤ −20% Green ✅ Low — keep monitoring
➡️ Stable −20% to +20% Yellow 📋 Medium — plan a fix

Each card includes a sparkline mini-chart, previous vs current failure count, urgency label, and a recommended action. Cards are sorted by urgency (increasing first), then by count.


Root Cause Analysis

Full-width section · 2-column layout

Platform-specific root cause detection — up to 10 cards total, sorted by impact.

Detection methods:

  • HTTP error code analysis (500 → Backend Error, 503 → Overload, 429 → Rate Limit, etc.)
  • Stack trace extraction and comparison
  • Keyword combinations (pool + exhausted → Connection Pool Issue)
  • Weighted similarity scoring (stack traces weighted 1.5×)

Confidence levels:

Range Color Interpretation
90–100% Green High confidence — fix it
70–89% Orange Likely correct — investigate
60–69% Yellow Worth checking

Each card provides evidence bullets, affected teams, detected error codes, and two action buttons: View X Tests (filters the main list) and Create Ticket.


Filtering & Search

Filters available:

  • Platform — All / Android / iOS / Web
  • Team — All teams or a specific one
  • Date Range — Start + end date pickers with Apply button

All filters combine. Results update in real time.

Search (top-right) matches across test name, failure description, platform, and team. Partial and case-insensitive matching is supported.


Test Detail Modal

Click any individual test to open a full-detail modal:

  • Test name, ID, platform, team, and execution date
  • Full failure description and stack trace
  • Detected error codes

Close with ×, click outside the modal, or press Escape.


Data Import

CSV Format

test_case, failure_description, platform, date, team_name, test_id, stack_trace

Column aliases accepted:

Field Accepted column names
Test name test_case, testcase, test
Error failure_description, description, error
Platform platform
Date date (YYYY-MM-DD)
Team team_name, team
ID test_id, id
Stack trace stack_trace, stacktrace (optional but recommended)

Including stack traces significantly improves root cause confidence.


Technical Reference

Stack: Pure HTML5 + CSS3 + ES6 JavaScript · No framework · No build tools · No external dependencies

Browser support: Chrome 90+, Firefox 88+, Safari 14+, Edge 90+, iOS Safari, Chrome Mobile

Performance targets:

Operation Target
Load 2,000 tests < 1 second
Filter / search < 100ms
Chart render < 200ms
Modal open < 50ms
Page change < 100ms

Algorithm complexity:

Algorithm Complexity Used for
Levenshtein distance O(m × n) Error message similarity
Stack trace matching O(n) extract + O(1) compare Call stack matching
Advanced grouping O(n²) worst / O(n log n) avg Test clustering

Use Cases

Daily standup — Load latest data, check Trend Analysis for 📈 increasing issues, assign 🚨 critical items.

Sprint planning — Review Root Cause Analysis sorted by impact, estimate fixes, prioritize by ROI.

Incident response — Filter by incident date range, identify root cause with high confidence, verify fix with Trend Analysis.

Regression detection — Monitor daily trends, watch for sudden 📈 increases, correlate with recent deployments.

Executive reporting — Use before/after Trend Analysis to demonstrate improvement and ROI of quality initiatives.


Pro tip: Start with Load Demo Data to explore all features before uploading your own CSV.

About

No description, website, or topics provided.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

 
 
 

Contributors

Languages