DevOps Engineer | Cloud Engineer | Learner
Our production cluster had free CPU everywhere — and pods stuck Pending because the subnets ran out of IP addresses. Why the VPC CNI eats IPs, the /24 → /20 subnet migration that bought us headroom, and the real fix we're heading towards.
Published: Jul 2026AWS just shipped EKS version rollback — revert a control plane upgrade to the previous minor version within 7 days, at no cost. Here's how it works, what "in-place only" means, and how it changes your upgrade strategy.
Published: Jul 2026Migrating from AMD to ARM (Graviton) nodes, we needed the frontend to stay on AMD. A PriorityClass alone wasn't enough — here's the full combination of nodeSelector, PriorityClass, and taints/tolerations that actually works.
Published: Jun 2026Most teams track CIDRs in a spreadsheet until two VPCs overlap and peering breaks. Here's how Amazon VPC IPAM plans, tracks, and monitors your IPs across accounts and regions — and even trims your public IPv4 bill.
Published: Jun 2026I scaled our dev EKS cluster to zero overnight to save cost. The first morning it came back, every pod pulled its image at once and the cluster threw a "QPS limit exceeded" error. Here's the root cause and the fix.
Published: Jun 2026A real-world case study on reducing billing by moving VPC Flow Logs to S3 in Parquet format.
Published: Feb 2026Understanding the journey of a DNS query from your browser to the server. Essential knowledge for troubleshooting connectivity.
Published: Feb 2026