Engineering retros, not generic tutorials.

Schemas I've owned in production — an 85-model construction SaaS, an identity platform, a staking backend, a marketplace API — taught me a short list of rules I no longer negotiate on. Snapshots for money, triggers for audit, constraints for invariants, and the isolation level nobody changes.

The checklist looks short: network, cluster, nodes, ECR, deploy. Each line hides a day of work. A field guide from taking a real app to an EKS cluster — nine subnets, Pod Identity, managed services, and the manifest that finally ships it.

An open-source Go CLI that scans an AWS account for idle and orphaned resources and estimates the monthly cost — the Trusted Advisor checks you shouldn't need a Support plan for. Here's the origin story.

Self-hosting traces, logs, and metrics on a single EC2 instance with Grafana, Alloy, Tempo, and Loki — how the pieces fit, the push-vs-pull split that trips people up, wiring it for a Go backend, and the caveats nobody mentions.

How blue-green deployments and canary releases actually achieve zero downtime — the traffic mechanics, the database problem nobody mentions, the tooling that runs them, and when not to bother.

A working playbook for auditing a production cloud stack — twelve layers from network perimeter to CI/CD pipeline, with verification commands, severity ratings tied to remediation SLAs, and the trade-offs senior engineers make in practice.

Start at the bit, walk up through CIDR, subnets, Security Groups, and DNS. Animated diagrams for the hierarchy and the TCP handshake — plus a debugging toolbox for when things don't connect.

A re-reading of Fowler, Martin, and the Gang of Four — what each book was reacting to, what holds up after twenty years, and where the cargo cult has drifted from the source.

Agentless config management explained — the mental model, the execution flow, and the commands you'll actually use.

Notes on shipping in public, building a portfolio, and what I plan to write about here.