Skip to main content
InfraGap.com Logo
Home
Getting Started
Core Concept What is a CDE? How It Works Benefits CDE Assessment Getting Started Guide Inner Loop vs Outer Loop Environment Drift Local vs Cloud CDEs for Startups
AI & Automation
AI Coding Assistants Agentic AI AI-Native IDEs Agentic Engineering AI Agent Orchestration AI Governance AI-Assisted Architecture Shift-Left AI LLMOps Autonomous Development AI/ML Workloads CDEs for Data Science GPU Computing
Agent Infrastructure
Agent Experience (AX) Agent Egress Control Computer Use Agents Agent Evals Agent Runbooks Agent Client Protocol AGENTS.md MCP Servers Git Worktrees Kubernetes Agent Sandbox Agent Fleets Agent Identity Prompt Injection Defense Agent Observability Context Engineering AI Code Review Bottleneck Headless Agents in CI Spec-Driven Development Agent Readiness Code Provenance
Implementation
Architecture Patterns DevContainers Advanced DevContainers Language Quickstarts IDE Integration CI/CD Integration Platform Engineering Developer Portals Container Registry Multi-CDE Strategies Remote Dev Protocols Nix Environments Hermetic Builds OpenTofu for CDEs Kubernetes Development
Operations
Performance Optimization High Availability & DR Disaster Recovery Monitoring Capacity Planning Multi-Cluster Development Troubleshooting Runbooks Ephemeral Environments Sandbox Environments Workspace Snapshots Database Branching
Security
Security Deep Dive Zero Trust Architecture Secrets Management Vulnerability Management Network Security IAM Guide Supply Chain Security Air-Gapped Environments AI Agent Security MicroVM Isolation Compliance Guide EU AI Act Cyber Resilience Act Data Residency Governance
Planning
Pilot Program Design Stakeholder Communication Risk Management Migration Guide Cost Analysis FinOps GreenOps Vendor Evaluation Training Resources Developer Onboarding Team Structure Platform Maturity Model Open Source CDEs AI Productivity Paradox Build vs Buy DevEx Metrics Productivity Engineering Industry Guides CDEs for Healthcare CDEs for Financial Services CDEs for Government Edge Development WebAssembly in CDEs
Resources
Tools Comparison State of CDEs 2026 Isolation Decision Tool Template Library
Learning Paths
All Paths Platform Engineer Security and Compliance Engineering Manager
Vendor Reviews
GitHub Codespaces Coder Ona Google Workstations Microsoft Dev Box Okteto Eclipse Che DevPod Daytona E2B
Head to Head
Coder vs Codespaces Coder vs Ona Ona vs Codespaces Coder vs Okteto Self-Hosted vs Managed E2B vs Daytona CDE Market Guide CDE vs Alternatives Case Studies Lessons Learned Glossary FAQ Sources & Citations

Platform Engineer Learning Path

Six stages through the pages on this site, ordered so that each one arrives after the thing it depends on - from the inner loop you are optimizing to the recovery plan you will eventually need

A route through existing material. Every entry is a page you could find on your own; the value here is the order and the reason for it.

Who This Is For

This path assumes you will operate the thing, not just choose it. That changes what is worth reading and in what order. It starts with the developer loop you are trying to improve, because every architectural decision later is judged against whether it made that loop better or worse, and a platform built without that reference point tends to optimize for the operator instead.

By the end you should be able to write an environment definition and defend its trade-offs, describe how a workspace is provisioned and torn down in your own stack, choose an isolation boundary deliberately, instrument the platform so a regression is visible before a developer reports it, and say what happens when the control plane is unavailable. Six stages, each a reasonable stopping point.

Stage One: the Problem You Are Actually Solving

Ends with you able to state, in one sentence, which loop you are shortening and what you are trading for it.

Inner Loop

First, because it is the measurement everything else is judged against. Read it before any architecture page or you will optimize the wrong number.

Environment Drift

The specific failure this category exists to remove. It names the symptom your stakeholders will describe to you in vaguer words.

Local vs Cloud Development

Now the honest comparison, including what gets worse. Reading it third means you already know what the loop and the drift cost you.

Core Concept

Vocabulary, deliberately after the problem statement. Definitions land better when you already have something to attach them to.

How It Works

The request path end to end. Close this stage with the mechanism, so the next stage is about choices rather than mysteries.

Stage Two: Environment Definitions

Ends with you able to write a definition for a real repository and explain where the specification stops and your platform starts.

DevContainers

The format most platforms consume, field by field. Everything downstream assumes you can read one of these fluently.

Advanced DevContainers

Features, multi-container stacks, and lifecycle hooks. This is where start time and isolation quietly get decided.

Quickstarts

Concrete configurations to test your understanding against, now that the fields mean something rather than being incantations.

Nix Environments

The alternative model, read after the mainstream one so the comparison is concrete rather than ideological.

Hermetic Builds

What reproducibility costs once you try to enforce it. It reframes both previous pages, which is why it comes last here.

Template Library

The copyable artifacts, useful only now that you can tell which caveat applies to you.

Stage Three: Architecture and Provisioning

Ends with you able to draw your own control plane and name every component somebody has to operate.

Architecture

The control plane and data plane split. Every operational decision in stage six traces back to this diagram.

OpenTofu

Templates as infrastructure code. Comes after architecture because a template is only meaningful once you know what it provisions into.

Container Registry

Where images come from, which becomes a dependency you own and a single point of failure you inherited without noticing.

Ephemeral Environments

The lifecycle model. Read it before snapshots, because it defines what disposable is supposed to mean.

Workspace Snapshots

The state that has to survive a stop. This is the compromise that makes disposable workspaces tolerable in practice.

Database Branching

The hardest state problem, deliberately last. It is usually the reason a clean ephemeral model does not survive contact with a real application.

Stage Four: Kubernetes and the Isolation Boundary

Ends with you able to choose an isolation boundary on purpose and say what it does not protect against.

Kubernetes Development

The most common substrate for workspaces, and the scheduling behavior you will be debugging.

Multi-Cluster

What changes when one cluster is not enough. Read it here rather than later, because it constrains the availability design in stage six.

Sandbox Environments

Isolation as a first-class requirement rather than a side effect of the scheduler.

MicroVM Isolation

The stronger boundary and what it costs in start time. This is the choice a permissive pod template quietly made for you.

Kubernetes Agent Sandbox

The same problem when the occupant is an agent rather than a person, which raises the bar on everything above.

GPU Computing

The workload that breaks most of the assumptions in this stage, which is why it belongs at the end of it.

Stage Five: Platform Practice

Ends with you able to position this platform inside a wider developer platform rather than as an isolated tool.

Platform Engineering

The discipline this work sits inside. It arrives now because the practice makes more sense once you have built something.

Platform Engineering Maturity

An honest read on where you currently are, which is more useful after stages two through four than before them.

Developer Portals

How a workspace becomes discoverable rather than something people ask about in chat.

CI/CD Integration

Reusing one environment definition across development and pipelines, which is where the drift work in stage one finally pays off.

Headless Agents in CI

The same pipelines with no human at the keyboard, and the capacity implications that come with it.

Stage Six: Operate It

Ends with you able to describe what breaks, how you would see it, and what you would do at three in the morning.

Monitoring

Instrumentation first, because everything that follows depends on being able to see it happen.

Agent Observability

The signals a human workspace never needed - token spend, tool calls, and runs that do nothing expensively.

Performance Optimization

Now that you can measure, this is where the inner loop from stage one gets its actual improvement.

Capacity Planning

Sizing from observed demand rather than from a vendor's calculator. It requires the data the previous three pages produce.

High Availability

What happens when the control plane is not there. Read it against the architecture you drew in stage three.

Disaster Recovery

The harder version of the same question, where the platform database is the thing you lost.

Runbooks

Turning all of the above into something a colleague can execute without you awake.

Troubleshooting

The symptom-first index, kept last as a reference rather than a lesson.