Blog
Notes on DevOps, infrastructure, and automation
CPU Starvation Without Hitting the Limit: A Kubernetes Latency Incident
Our Next.js pods never came close to their CPU limit. Usage sat right at the request value and the app was still dying with ECONNRESET. Raising the CPU request from 100m to 500m fixed it — here's why the limit was never the problem.
5 kubectl Aliases That Save Me Hours Every Week
If you're running Kubernetes daily, typing full kubectl commands gets old fast. Here are the 5 aliases I added to my .zshrc that I genuinely can't work without anymore.
PgBouncer for Production Applications
Discovered 1176 idle PostgreSQL connections in production. Here's how PgBouncer as a centralized connection pooler brought that down to 10–20 active connections and saved us from a costly headache.
Kubernetes Best Practices for Production Clusters
Lessons learned from running Kubernetes at scale - resource limits, namespace strategy, and observability patterns that actually hold up in production.
Designing CI/CD Pipelines at Scale
How to build CI/CD pipelines that stay fast and reliable when hundreds of teams are shipping through them every day.