Aaron Aldrich - DevOpsDays NYC 2018 Ignite
summary
Argues that optimizing for mean time to resolution beats stretching mean time between failures, and that resilient systems recover from unknowns rather than pretending to prevent every fault. Describes tools and practices used at elastic across 30 countries and 19 different languages, like using / zoom from slap to move from text to video, adopting blameless postmortems, and using non-violent communication to surface observable facts, feelings, needs, and requests.
AI-written summary · about
chapters (8)
- 0:00Start
- 1:46unpaid debt leads to catastrophic failure
- 2:01there is no root cause to failure in complex systems relationships
- 2:16blamelessness (just culture) is required for improvement
- 2:31experience with failure is necessary
- 3:16What you're saying is 'X', what I'm hearing is 'Y'.
- 3:46assume good intent
- 5:23Recording and Production Confreaks
related talks
40mDistilled Abnormal DevOps Iterations
34mToward continuous resilience
30mThe last in the line: a talk about DevOps and Disasters
27mRyn Daniels at The Lead Developer New York 2017
10mLightning Talk: A Culture of Resiliency through Chaos
47mDevOps is for Product Engineers, too.
15mMaintaining reliable systems
32mDistributed DevOps