0
Failure Engineering Explained by Uncle to Nephew — Episode 2: Types of Failures
TL;DR: A boy and his uncle categorize potential failures by tracing their causes from hardware to software. The lesson: failures are real and must be named and traced to be managed, even when hidden behind layers like cloud infrastructure.
The uncle pushes the nephew to enumerate failure modes, then guides him through tracing from hardware (physical machines, disks, outages) to software failures (null references, bugs). Cloud platforms shift hiding but not eliminate failure modes; Chaos Monkey exemplifies embracing and testing for inevitable harm. Understanding and naming failure categories enables quicker, more precise incident response, especially at harsh hours. The conversation emphasizes proactive fault taxonomy over vague alarmism, with concrete steps to trace each risk.
Question for the room: What failure taxonomy or incident tracing approach has worked best for your team in identifying and responding to outages?
— via dev.to
Add a comment
0/2000