Nitin Agrawal·Apr 23Kafka Producer Buffer Memory: The Silent OOM Killer in Spring Boot AppsOur Spring Boot service ran fine for three weeks.
Nitin Agrawal·Mar 25Your Codebase Isn’t AI-Ready — Here’s How We Scored Ours Using FoundryWe let an AI assistant refactor a service last Tuesday. It confidently renamed half the variables, hallucinated a dependency that doesn’t…
Nitin Agrawal·Mar 7Every Feature Needs One Thing Before Release: AlertsImagine shipping a feature you’re proud of — polished UI, solid unit tests, code reviewed twice — only to find out three days later that it…
Nitin Agrawal·Mar 3Why P99 Latency Is the Only Metric That Actually MattersYour monitoring dashboard shows average response time: 42ms. Green across the board. The team is happy. The sprint review goes smoothly…
Nitin Agrawal·Mar 2The One Thing Every Great Codebase Does DifferentlyI’ve shipped features at 2am, inherited systems that made me question my career choices, and refactored code I wrote myself. Here’s what…
Nitin Agrawal·Feb 24Using Compression to Reduce AWS RDS CostsLast month, I watched our AWS bill climb past $8,000 for RDS alone. Our startup’s database had grown from a modest 100GB to over 2TB, and…
InTowards DevbyNitin Agrawal·Feb 23Your MySQL Database Isn’t Growing — It’s HoardingPicture this: It’s 2 AM. Your on-call alert fires. The database disk is at 98%. You SSH in, heart pounding. You start digging — and you…
InTowards DevbyNitin Agrawal·Feb 6Our Kafka Payloads Randomly Stopped Sending — The Real Culprit Was Producer Buffer MemoryThe customer support ticket said: “My order confirmation never arrived.”A response icon1A response icon1
Nitin Agrawal·Feb 3Your System Isn’t Slow — Your Architecture Is Lying to YouOur API was “slow.” P95 latency: 850ms. Unacceptable.
InTowards DevbyNitin Agrawal·Feb 2Why Go Wastes Memory on Purpose (And Why That’s a Good Thing)My Go application was using 2.3GB of RAM.