Retrospective Analysis: From 5 Monolithic Pipelines to 4-Layer Architecture

Refactoring Before After
What I did Transformed 5 monolithic pipelines (41% duplication) into 4-layer architecture with centralized libraries
How it emerged Felt the pain - organized by domain naturally - patterns emerged
Post-analysis Mapped my solution to industry concepts (Software Smells, DDD patterns, GoF Design Patterns)
Documentation Problem analysis, pattern identification, architecture rationale

Note on 37% vs 41%: If you came from my resume, you may have seen "37% duplication." That figure was from an earlier baseline (54479b2, 2025-02-21). After detailed commit history analysis, 74fc356 (2025-03-20) was identified as the accurate "before refactoring" state, showing 41% duplication. The increase itself demonstrates the architecture's High Viscosity problem.

Purpose

This portfolio is a post-refactoring analysis of intuitive problem-solving done during VARLab Co-op.

What I Did What This Document Does
Refactored 5 monolithic pipelines into 4-Layer Architecture Classify and define those intuitive fixes using industry conventions
Designed 3-Level Logger System Analyze why the design worked and how it integrates with the architecture

Goal: Transform experience-based understanding into systematic, reusable knowledge

If I Had This Analysis Then...

Before (During Refactoring)

When I proposed this refactoring to my supervisor, I could only say:

"The code is messy. There's a lot of copy-paste. It's hard to maintain. I think we should restructure it."

I knew something was wrong. I felt the pain every time I had to modify multiple files for a single change. But I couldn't articulate why it was problematic or how bad it actually was.

After (With This Analysis)

If I had this analysis back then, I could have presented:

"We have 41% code duplication across 5 pipelines. A single change to sendBuildStatus requires modifications in 18 locations across 6 files. We're making 113 external calls to 12 different systems with no abstraction layer, making the code untestable. In the past month alone, duplication increased from 37% to 41%, showing the architecture's High Viscosity is causing duplication to grow naturally."

Instead of "it feels messy," I could have shown concrete numbers. Instead of "it's hard to maintain," I could have explained Shotgun Surgery with specific examples. Instead of "we should fix it," I could have demonstrated the ROI degradation where 10-minute tasks were taking hours.

The Difference

Aspect Before After
Problem Description "Code is messy" "41% duplication, 18-location Shotgun Surgery"
Justification "I feel it's hard to maintain" "Duplication increased 4% in 1 month (High Viscosity)"
Scope "A lot of copy-paste" "113 external calls, 12 types, untestable"
Urgency "We should fix it sometime" "10-20 min tasks - hours, ROI degradation"

This analysis transforms intuition into evidence. It provides the vocabulary and metrics needed to communicate technical debt to stakeholders who may not see the code daily.

What Would You Like to Know?

🔍 Problem Analysis

Q: Want to see how I analyzed a single file?

Analysis flow: Define SRP criteria - Analyze responsibility per function/module - Identify Software Smells

Q: Want the overall problem analysis summary?

Q: Want the detailed analysis?

🛠️ Solution

Q: Want to see the final architecture and solution overview?

Q: Want to know which design principles were applied?

Q: Want the step-by-step implementation process?

📄 Supplement

Q: Want the DDD perspective analysis?

Q: Want to see the Pull Request documentation?

Project Overview

Background

Timeline

Milestone Date Commit Repository
Initial commit (Windows) 2023-12-17 e9d8028 devops-jenkins-windows
Migration to Linux 2025-01-07 5a59d45 devops-jenkins-linux
Baseline (Before refactoring) 2025-03-20 74fc356 devops-jenkins-linux
Technical Debt Period ~14 months
Refactoring complete 2025-05-12 ff74ac8 devops-jenkins-linux

Repository started in Windows environment and migrated to Linux. Initial commit (2023-12-17) - Baseline (2025-03-20) = ~14 months of accumulated Technical Debt

Document Structure

retrospective-analysis-from-5-monolithic-pipelines-to-4-layer-architecture/
│
├── README.md                           # ← This document (entry point)
│
├── problem-analysis/                   # Problem Analysis
│   ├── problem-analysis-overview.md    # Problem summary (★ Start here)
│   ├── detailed-analysis.md            # Synthesized findings
│   ├── architecture-smells-analysis.md # System-level analysis
│   ├── DRY-violation-analysis.md       # Cross-file duplication
│   │
│   ├── 01-generalHelper/               # Per-file analysis
│   │   ├── 01-srp-violation-analysis.md
│   │   ├── 02-software-smells-analysis.md
│   │   └── 03-design-smells-symptoms.md
│   ├── 02-jsHelper/
│   ├── 03-unityHelper/
│   ├── 04~08-Jenkinsfiles/
│   │
│   └── pipeline-sequence-diagrams/     # Pipeline flow analysis
│       ├── domain-mapping.md
│       └── dlx-ci.md, dlx-cd.md, js-ci.md, js-cd.md, jenkins-ci.md
│
├── problem-solving/                    # Solution
│   ├── solution-overview.md            # Entry point for solution docs
│   ├── solution-by-layer.md            # 4-Layer Architecture analysis
│   ├── solution-by-feature.md          # Logger System, Shell Libraries
│   ├── logger-system-design-integration.md  # Logger design document
│   ├── domain-driven-analysis.md       # DDD analysis (supplement)
│   ├── pull-request-documentation.pdf  # PR attachment
│   └── phase-0~4/                      # Step-by-step changelogs
│
└── resources/                          # Images
    └── architecture-overview.png, stage-logger-*.png

Reference