20+ years of building systems that outlast the chaos
summary
Argues that teams must scale down before you scale up and learn kernel, memory, and app fundamentals instead of buying complexity; replaces fragile deploy scripts by packaging apps as RPMs so Tomcat plus JDK install with one dnf command; demo shows the RPM-based install works, eliminates brittle scripts, lets product teams deploy via Jira with automatic rollback; cites CockroachDB running on a laptop as a design litmus and an EBT payment outage to show the real human cost of unreliable systems.
AI-written summary · about
chapters (13)
- 0:00Introduction
- 1:12Human expertise in systems
- 2:37Scaling down before scaling up
- 3:42Enterprise solutions illusion
- 4:46Small wins matter
- 7:58First job and deployment chaos
- 11:32Unreliable deployments fail
- 13:19Human impact in systems
- 15:26Learning Java and simpler deployments
- 17:55Building RPMs and reducing complexity
- 19:19Power of simple commands
- 20:25Talking about impact
- 21:08Fundamentals over AI hype
related talks
01h 02mTools for simplified large scale systems
53mChange your tools, change your outcome
05mAdventures In Enterprise: Lessons Learned Promoting DevOps In A Legacy Organization
31mRadical Ideas Enterprises Can Learn From The Cloud
14mAntifragility or taming the hydra
41mDystopia as a Service
01h 02mPractical Chaos Engineering breaking things on purpose to make them more resilient against failure
50mKeynote - More than two decades in shell and ops