Yoshik Karnawat

SRE @PhonePe

Digest

#17 Uptime Sync: AI Coding's Junior Dev Killer, RFCs as the Ultimate Engineering Skill

May 2, 2026

•

2 min read

#17 Uptime Sync: AI Coding's Junior Dev Killer, RFCs as the Ultimate Engineering Skill

Microsoft's AI agent career warnings, Istio's registry migration fallout, and the chaos engineering practices that actually ship to production

Yoshik Karnawat

May 1, 2026

•

1 min read

Cracking SRE

Access the Guide

Yoshik Karnawat

Networking

Apr 30, 2026

•

5 min read

TCP is Dead

Why QUIC on UDP is the web's fix

Yoshik Karnawat

Digest

#16 Uptime Sync: MySQL Repo, AI Cognitive Overload, and AWS Incident Lessons From 3000 Outages

Apr 26, 2026

•

2 min read

#16 Uptime Sync: MySQL Repo, AI Cognitive Overload, and AWS Incident Lessons From 3000 Outages

NPM supply chain risks, monorepo misconceptions, and the real AI productivity gains hiding in plain sight

Yoshik Karnawat

Digest

#15 Uptime Sync: QUIC, K8s API Governance, and Infrastructure Lessons That Actually Ship

Apr 18, 2026

•

2 min read

#15 Uptime Sync: QUIC, K8s API Governance, and Infrastructure Lessons That Actually Ship

Anthropic's Claude Design pivot, the engineering incentives that reward complexity over simplicity, and why serverless is quietly losing to single powerful servers

Yoshik Karnawat

System Design

Apr 14, 2026

•

23 min read

Consistent Hashing

The Algorithm Behind Every Distributed System You Use

Yoshik Karnawat

Digest

#14 Uptime Sync: GitOps, AI Agents, and Kubernetes at Scale

Apr 11, 2026

•

4 min read

#14 Uptime Sync: GitOps, AI Agents, and Kubernetes at Scale

Anthropic's secret project, the npm supply chain attack you should've caught, and the Kubernetes cost cut hiding in your instance type

Yoshik Karnawat

Digest

#13 Uptime Sync: Axios Hack, AWS Multi-AZ Assumptions Broken by War, Wix Migrated 1000 MySQL Servers

Apr 4, 2026

•

4 min read

#13 Uptime Sync: Axios Hack, AWS Multi-AZ Assumptions Broken by War, Wix Migrated 1000 MySQL Servers

Iran conflict outage, Graviton zero-downtime migration, Redpanda Cloud architecture, and the AWS logs you miss during incidents

Yoshik Karnawat

Apr 1, 2026

•

6 min read

Understanding Kubernetes Scheduler

Filter, Score, Bind: and why the first phase determines everything that follows

Yoshik Karnawat

Digest

#12 Uptime Sync: DoorDash's Service Mesh at 80M RPS, GitLab Deployed 12 Times Daily, and AI Agent Misconceptions

Mar 28, 2026

•

4 min read

#12 Uptime Sync: DoorDash's Service Mesh at 80M RPS, GitLab Deployed 12 Times Daily, and AI Agent Misconceptions

IP overlap routing, GitHub Enterprise search HA, storage design trade-offs, and S3 as a database substrate

Yoshik Karnawat

System Design

Mar 23, 2026

•

6 min read

A Crash Course On High Availability

Your App Being "Up" and Your App Being "Available" Are Not the Same Thing

Yoshik Karnawat

DevOps

Your Docker Image Isn't Fat Because of Your Code

Mar 10, 2026

•

9 min read

Your Docker Image Isn't Fat Because of Your Code

The average production Node.js image sits at 900MB. It should be under 150MB. Here's the architectural reason why.

Yoshik Karnawat

Digest

#11 Uptime Sync: GPT-5.4 Launch, Yelp's AI Assistant, and GPU Health at 20K Scale

Mar 7, 2026

•

5 min read

#11 Uptime Sync: GPT-5.4 Launch, Yelp's AI Assistant, and GPU Health at 20K Scale

Multi-agent architectures, production AI agent skepticism, infrastructure audit automation, and classical ML inference breakthroughs

Yoshik Karnawat

Digest

#10 Uptime Sync: Docker Builds in Seconds, Netflix LLM Post-Training, and Uber's Access Control Reimagined

Mar 2, 2026

•

5 min read

#10 Uptime Sync: Docker Builds in Seconds, Netflix LLM Post-Training, and Uber's Access Control Reimagined

Build optimization breakthroughs, runtime security hardening, database partition strategies, and production-tested infrastructure tools

Yoshik Karnawat

Cloud

Feb 24, 2026

•

7 min read

The Secret Behind AWS S3

The "11 nines" everyone quotes is not an uptime promise

Yoshik Karnawat

#9 Uptime Sync: Claude Code Security, Microservices and Databases, and GitOps Lessons Learned

Feb 21, 2026

•

4 min read

#9 Uptime Sync: Claude Code Security, Microservices and Databases, and GitOps Lessons Learned

AI coding security, database architecture patterns, promotion frameworks, and self-hosted infrastructure tools worth knowing

Yoshik Karnawat

#8 Uptime Sync: GitHub’s Feb 9 Incident, Gemini 3 Deep Think & RDMA

Feb 15, 2026

•

4 min read

#8 Uptime Sync: GitHub’s Feb 9 Incident, Gemini 3 Deep Think & RDMA

Platform reliability lessons, new model research, security pitfalls in “vibe coding,” and systems-level performance primitives

Yoshik Karnawat

System Design

Feb 10, 2026

•

10 min read

YouTube Architecture

How YouTube grew to 2 billion users by bending MySQL to their will

Yoshik Karnawat

Digest

#7 Uptime Sync: GPT-5.3 Codex, Claude Opus 4.6 & Netflix's Real-Time Graph

Feb 8, 2026

•

4 min read

#7 Uptime Sync: GPT-5.3 Codex, Claude Opus 4.6 & Netflix's Real-Time Graph

New AI model launches, credential leak research, and distributed systems architecture at internet scale

Yoshik Karnawat

Digest

#6 Uptime Sync: Kubernetes 1.35, PostgreSQL at 800M Users & TCP Debugging

Feb 2, 2026

•

4 min read

#6 Uptime Sync: Kubernetes 1.35, PostgreSQL at 800M Users & TCP Debugging

Scaling 800M users, production security stories, and network debugging patterns that actually work

Yoshik Karnawat

Networking

Jan 27, 2026

•

7 min read

HTTP/3 Doesn't Just Speed Things Up

It Abandons TCP Entirely

Yoshik Karnawat

Jan 19, 2026

•

6 min read

How SSO Actually Works

One login, dozens of apps. Here's the mechanism behind it.

Yoshik Karnawat

Digest

#5 Uptime Sync: Zero-Downtime Elasticsearch Upgrades, AI Realities & Kubernetes Mental Models

Jan 16, 2026

•

5 min read

#5 Uptime Sync: Zero-Downtime Elasticsearch Upgrades, AI Realities & Kubernetes Mental Models

Major version upgrades without downtime, honest takes on AI limitations, distributed system architecture patterns, and production-grade security tooling

Yoshik Karnawat

Cloud

Jan 12, 2026

•

8 min read

Under-the-hood of Load Balancers

Most 502s don't come from failed backends. They come from behaviors you never configured.

Yoshik Karnawat

Digest

#4 Uptime Sync: BGP Route Leaks, GitHub Platform Engineering & Terraform Rollback Patterns

Jan 8, 2026

•

4 min read

#4 Uptime Sync: BGP Route Leaks, GitHub Platform Engineering & Terraform Rollback Patterns

Route leak forensics in production networks, how platform teams solve infrastructure at scale, safe infrastructure rollback strategies, and AI coding agents that actually ship code

Yoshik Karnawat