惯性聚合 高效追踪和阅读你感兴趣的博客、新闻、科技资讯
阅读原文 在惯性聚合中打开

推荐订阅源

J
Java Code Geeks
F
Fortinet All Blogs
Martin Fowler
Martin Fowler
M
MIT News - Artificial intelligence
G
Google Developers Blog
P
Proofpoint News Feed
Recent Announcements
Recent Announcements
MyScale Blog
MyScale Blog
D
DataBreaches.Net
Stack Overflow Blog
Stack Overflow Blog
月光博客
月光博客
爱范儿
爱范儿
罗磊的独立博客
腾讯CDC
Hugging Face - Blog
Hugging Face - Blog
博客园 - 叶小钗
Vercel News
Vercel News
酷 壳 – CoolShell
酷 壳 – CoolShell
B
Blog
C
Check Point Blog
美团技术团队
宝玉的分享
宝玉的分享
Microsoft Security Blog
Microsoft Security Blog
OSCHINA 社区最新新闻
OSCHINA 社区最新新闻

DEV Community

Authentication Security Deep Dive: From Brute Force to Salted Hashing (With Java Examples) Why AI Systems Don’t Fail — They Drift Spilling beans for how i learn for exam😁"Reinforcement Learning Cheat Sheet" I Replaced Chrome with Safari for AI Browser Automation. Here's What Broke (and What Finally Worked) How Python Borrows Other People's Work The $40 Architecture: Processing 1 Billion API Requests with 99.99% Uptime Vibe Coding: A Workflow Guide (From Zero to SaaS) Most webhook security guides protect the wrong side. The scary part is delivery. Headless CMS for TanStack Start: Build a Blog with Cosmic EU Age Verification App "Hacked in 2 Minutes" — What Actually Happened Comfy Cloud’s delete function does not actually remove files Running AI Models on GPU Cloud Servers: A Beginner Guide Event-driven media intelligence with AWS Step Functions and Bedrock I scored 500 AI prompts across 8 quality dimensions — here's what broke How to Call Google Gemini API from Next.js (Free Tier, No Backend Needed) The Portal Protocol: Reclaiming Human Connection in the Age of AI How to Fix Your Team's Scattered Knowledge Problem With a Self-Hosted Forum Intro to tc Cloud Functors: A Graph-First Mental Model for the Modern Cloud Designing Multi-Tenant Backends With Both Ownership and Team Access I Built a Neumorphic CSS Library with 77+ Components — Here's What I Learned PostgreSQL Performance Optimization: Why Connection Pooling Is Critical at Scale Cómo construí un SaaS multi-rubro para gestionar expensas en Argentina con FastAPI + Vue 3 🚀 I Built an Ethical Hacking Scanner Tool – Open Source Project I Replaced /usage and /context in Claude Code With a Single Statusline A Pythonic Way to Handle Emails (IMAP/SMTP) with Auto-Discovery and AI-Ready Design I Collected 8.9 Million Polymarket Price Points — Here's What I Found About How Markets Really Move EcoTrack AI — Carbon Footprint Tracker & Dashboard Everyone's Using AI. No One Agrees How. 5 self-hosted ebook managers worth trying in 2026 Building Your First AI Agent with LangChain: From Chatbot to Autonomous Assistant
If Your Server Is Bored, You’re Doing It Wrong
Mohit Kumar Kushwaha · 2026-06-03 · via DEV Community

Mohit Kumar Kushwaha

Not utilising the server to the fullest is a sin

It all started when a senior AEM Developer, after seeing his pipeline fail for the 10th time straight and wasted over 15 man hours just for a single Pull Request, reached out to me angrily complaining about the Maven Build time being too much for a Pull Request this small.

Context: We use AEM with Maven for Build check bundled with React & solid.js. Total Build time was 95 minutes for each iteration. Along with this, SonarQube & GitHub Advanced Security Scanning stretched the total time to 120 minutes.

Hell, yeah! This is frustrating. Even for me, to see my servers being exploited like this, for 2 hours straight.

Well, my initial reaction was “How dare he shout at me like this, as if I created this process and I am wasting his 2 hours of time”.

But then kicked in my instincts and I started talking with my own thoughts, “I didn’t setup this process, but i surely can improve it for better right?”

I analysed

  1. Project architecture and components of Build
  2. Server utilisation — cpu, memory, disk
  3. real-time resource monitoring and performance checks

Conclusion was, “We were the culprits all along”. because we were barely using the server power for Build check, Testing, Linting and SonarQube scanning. There are roughly 12000+ tests, and they were running Synchronously one by one. Can you believe it?

My runner Server is of size: c5a.4xlarge (16 vCPUs, 32 GiB of memory) and being used only 5%.

We decided to tweak the configurations. And after rounds of experimentation aiming for the perfect recipe. I found the proportions which were perfect.

They were

  1. Defining the heap size limits: 1024 –> 4096 MB of memory
  2. Parallel builds scaled to 1.5× CPU cores — on a 16-core machine, that’s 24 concurrent build threads.
  3. Reusing the forks — helps accelerate the JVM resources reusability for each test set.
  4. Tiered Compilation — Runtime profiling for compiling and warm up. Asynchronous module compilation, defying the dependency.

Drum Rolllllssssssssssss

29 minutes!!!! Yes, this combination of configurations brought down the total build time to 29 minutes. A reduction of ~60 minutes. Crazy enough.

What started as frustration quickly evolved into something crazy that my entire developer team loves now.

Most of the times, the processes are fine, but the systems are not, because we are not utilising them properly.

Utilise, what you pay for!!

Illustration of 2 server scenarios

P.S. 8 of my Runner Servers died(crashed) during the entire testing process in search of the right recipe of configurations.