Battles Wounds from a Cloud SRE.. And a few lessons
summary
Cloud SRE war stories focus on concrete fixes not platitudes: automated provisioning without lifecycle cleanup triggers runaway cost and a hidden 10,000 backup limit; a request to cut 20% of spend turns into a 91% cloud spend reduction after deleting orphaned resources; an AWS off‑by‑one bug deletes production instances so enterprise support and robust DR matter; a wrong database node restart causes irreversible data loss so teams must automate creation and destruction, enforce certificate renewal automation, practice outages, and publish blameless retrospectives.
AI-written summary · about
chapters (9)
- 0:00Intro
- 0:59Why do we need Site Reliability Engineers?
- 1:15What is the purpose of a Cloud SRE?
- 2:03Battles: Sprawl
- 7:25Battles: F.U.B.A.R.
- 10:32Battles: Lacking Preparation
- 15:05Winning the War: Solid Retrospectives
- 16:13Winning the War: Transparency
- 18:13What more information?
related talks
40mAvoiding Pain When Operating in the Cloud
43mWhere SRE Can Go Wrong
47mPersistent SRE Antipatterns: Pitfalls on the Road to Creating a Successful
24mHow to Destroy an SRE
39mPerils of the cloud
41mWays to optimize Cloud Disaster Recovery cost
45mToward a Culture Where Everyone Is Responsible for Outages
31mSeeing Through the Fog - Navigating the Security Landscape of a Cloud-First World