SRE and adapting to the uncertainties of military conflict
summary
Kessel Run SRE and DevOps identify a frontend/backend bottleneck that trips when traffic exceeds ~5× normal; during the Kabul evacuation the system faces about 10× normal sorties and the app crashes. Operators trace logs, allocate cloud compute and storage, push updates across an air‑gapped pipeline and fixed production in 12 hours, restoring load times to 2–3 seconds so planners continue operations and the airlift moves roughly 126,000 people.
AI-written summary · about
chapters (9)
- 0:00SRE in military ops
- 1:35Kessel Run’s legacy challenges
- 3:01Surge capacity planning
- 5:3112-hour outage response
- 7:56SRE’s crisis response value
- 9:43SRE for rapid incident resolution
- 11:09SRE adoption recommendations
- 12:14Rhythm and sports analytics
- 13:43Q&A
related talks
43mSRE From Inception to Global Scale
41mSRE by Influence, Not Authority: How the New York Times Prepares for Large Scale Events
30mGet Your Non-SREs Oncall Ready!
13mWhat To Do When SRE is Just a New Job Title?
30mShipping Through Chaos: Engineering Resilience Across Borders
45mClearing the Way for SRE in the Enterprise
16mSRE in Transition: From Startup to Established Business
38m"Disorganizing" Your SRE Organization