DevOps Field Guide - From Theory to Production

A practitioner's field guide for the journey from fundamentals to production-grade DevOps engineering.

What This Is

This repository operates on three levels:

1. Field Guide - Distilled concepts, patterns, and principles for each DevOps domain, written to be referenced in practice
2. Learning Journal - Honest reflections, surprises, and connections discovered along the way
3. Live Projects - A rolling project that evolves with each module, accumulating real infrastructure

Course Roadmap

#	Module	Focus	Status
00	Intro	DevOps philosophy, SDLC, team dynamics	Done
01	Linux	Command line, filesystem, permissions, scripting	Done
02	Python Basics	Syntax, data structures, scripting	Done
03	Git & GitHub	Version control, branching, collaboration	Done
04	Advanced Python	OOP, Flask, APIs, automation	Done
05	Networking	TCP/IP, DNS, HTTP, firewalls	Done
06	AWS	Cloud fundamentals, core services	Done
07	Docker	Containers, images, Compose	Done
08	CI/CD	Pipelines, automation, deployment strategies	Upcoming
09	Kubernetes	Orchestration, pods, services	In Progress
10	Helm	Package management for Kubernetes	In Progress
11	AWS Advanced	ECS, ECR, EKS	Upcoming
12	Terraform	Infrastructure as Code	Upcoming
13	Jenkins	Build automation, pipeline orchestration	Upcoming
14	Observability	Grafana, monitoring, dashboards	Upcoming
15	ArgoCD	GitOps, continuous delivery	Upcoming

Iron Rules

Principles collected throughout the course. These are non-negotiable truths earned through practice.

• "An application is not code." - it is a solution living inside an organization
• "Tech decisions are also organizational decisions."
• "If it works, don't touch it."
• "Any system too open will be breached."
• "What you don't need, don't use."
• "Always think worst case."
• "A requirement not defined correctly doesn't disappear - it comes back as bugs."
• "Bug in code = incident. Bug in data = crisis."
• "Code that works in dev is not necessarily production-ready."
• "insert_one() mutates your dict — handle _id before responding."
• "Credentials in code = credentials in git. Use .env from day one."
• "Circular imports? Extract shared resources to their own module."
• "kubectl create -f is imperative-with-file, not declarative. Declarative is kubectl apply."
• "spec.containers is required in every Pod manifest - it's a list, not a single object."
• "kubectl describe is your proof that actual state == desired state."
• "connection refused on the API server = cluster probably isn't running. Check docker ps first."
• "kubectl create does not write last-applied-configuration. Migrating to apply triggers a one-time warning, then the annotation is backfilled automatically."
• "Old ReplicaSets at 0 replicas are not junk - they are rollback history. Don't delete them manually."
• "Never assume your working directory. pwd + ls -la before every lab."
• "ReplicaSet counts every Pod matching its selector - including manually created Pods. Do not mix them casually."
• "Changing ReplicaSet template does not update existing Pods. Only future Pods use the new template."
• "pod-template-hash connects Deployment template, ReplicaSet name, and Pod names. Never set it manually."
• "Changing pod template (image, env, etc.) creates a new ReplicaSet. Changing replicas does not."
• "Rollback creates a new revision. It does not restore the old revision number."
• "kubectl scale is temporary. kubectl apply always wins."
• "Scaling replicas does not create a new rollout revision because pod template did not change."
• "ImagePullBackOff is diagnosed at Pod level, not at Deployment level. Use kubectl describe pod."
• "RollingUpdate keeps old Pods running until new Pods are healthy. A broken image does not kill availability."
• "Rollback is fast recovery. If the manifest still has the bad image, the next apply brings it back."
• "change-cause annotation is documentation, not action. It can lie. Check the actual image."
• "kubectl diff before kubectl apply is not optional."

This list grows with each module.

About the Author

David Rubin - AI Solutions Specialist transitioning to DevOps Engineering.

Currently bridging the gap between business needs and technical implementation, now building depth in infrastructure, automation, and production systems.

• GitHub: @Dude775
• Portfolio: system-portfolio.vercel.app