Revision · Foundations — Why DevOps Exists

Part 0 built the mental model that the rest of the book stands on. Its throughline is one question — what manual, error-prone step does this remove, and how does it make production safer? — and its punchline is that speed and stability are not a tradeoff but two sides of the same automation.

What this part covered

• The wall between Dev and Ops — Dev is measured on velocity, Ops on uptime, so they see the same deploy as progress versus risk; Conway’s Law means that split leaks into a brittle handoff seam.
• The cost of throwing it over the wall — a manual handoff ships the artifact and a wiki of steps but not the context, and failures like Knight Capital’s $440M loss show what one missed manual step does in production.
• Dissolving the wall — shared ownership (“you build it, you run it”) fixes the opposing incentives, and automating the handoff replaces the manual over-the-wall steps with a pipeline.
• Feedback loops and batch size — shorten the loop from change to feedback; small frequent changes have a tiny blast radius and instant diagnosis, while big rare ones are the real risk and feed the “doom loop.”
• Lead time and nested loops — the clock from commit to running in production is the lever (elite: minutes; low: weeks), and loops nest from local compile to CI to deploy to product.
• The four DORA metrics — a speed pair (deployment frequency, lead time) and a stability pair (change failure rate, time to restore) that act as a lie detector because you can’t game all four at once.
• CALMS, culture first — Culture, Automation, Lean, Measurement, Sharing, with blameless culture first: blame hides the very information that prevents the next failure, so postmortems target the system, not the person.
• The false tradeoff — DORA data shows elite teams cluster fast and stable; automation, small batches, and fast recovery make “move fast” and “recover fast” the same discipline, proven by Amazon’s deploy-every-11.6-seconds cadence.

The takeaway

Part 0 is the lens, not the toolbox: every tool later in the book is one more manual, error-prone step being automated away to make production both faster and safer. Hold onto the recurring question and the DORA scoreboard, because the next parts — Linux, networking, containers — are where those principles first meet real tooling and start to bite.