Lessons from Mission-Critical Systems in Fast-Moving Industries
summary
Argues that mission‑critical systems need autonomous teams, measurable practices and concrete tooling: deployable code with access to infra, peer‑reviewed PRs and metrics like PR decline and lines per change. Shows failures (SSL expiry, 15k alert emails, a server stuck in a basement) and the fixes that work — adding open telemetry and trace IDs, creating automated runbooks, and switching to queue-based ingestion that stops daily crashes. Claims automation as default, not a separate project.
AI-written summary · about
chapters (17)
- 0:00Introduction
- 1:49Mission-critical infrastructure
- 3:36Teams running mission-critical systems
- 7:49High demand vs healthy work
- 9:56Architecture and automation
- 12:46Resilient and observable systems
- 15:16End-to-end observability challenge
- 17:22Fast feedback loops
- 18:25Measuring performance
- 20:58Postmortems and improvement
- 22:23Shift reliability left
- 23:07Checklists and runbooks
- 25:12Processes versus culture
- 26:16Making behaviors tangible
- 28:04From fragile to antifragile
- 32:17Key takeaways
- 33:45Q&A
related talks
26mMission Critical Mindset Building a Culture of Resilience in DevOps Teams
43mGetting Your System into Production and keeping It There
27mTrusting Mission Critical Infrastructure: Building Reliable CD P...
32mTales from the DevOps Transformation Trenches
39mOops!… We F*ed It Up Again: The Art of Staying Calm When Systems Crash
29mDevOps Anecdotes
54mBridging the Gap CICD Mind the Gap
01h 39mPostOps: A Non-Surgical Tale of Software, Fragility, and Reliability