Database Reliability Engineering
summary
Book frames database work as reliability engineering and uses concrete examples from MySQL, PostgreSQL, MongoDB, Cassandra and a Facebook RocksDB rollout to show what actually changed. It claims autoscaling replicas is non-linear and error-prone, demonstrates fleet-scale MongoDB operations and rollback-tested migrations, and separates instrumentation-first observability vs monitoring. Recommendations include quantified error budgets, runbooks, deployment checklists and postmortems — many of which are shown to reduce incidents in production.
AI-written summary · about
chapters (10)
- 0:00Introduction
- 1:32Tell us about yourself
- 3:27Why is it hard to build agile evasion
- 4:41GPA job has shifted tremendously
- 6:55Island of expertise
- 8:09Observability
- 11:39Database Reliability Engineering
- 13:15DevOps Remediation
- 14:20Whats Next
- 17:17Business Outcomes
related talks
46mDatabase Reliability Engineering for MySQL
49mOnline Database Reliability, Performance, and Consistency Engineering
52mRocksDB Performance and Reliability Practics
01h 20mProblems with Relational Databases in 2020
44m50 Years of Choosing the Wrong Database: Lessons from History (and My Own Projects)
51mMyRocks deep dive and production deployment at Facebook
47mDatabase Resiliency
47mLessons Learned from Our Main Database Migrations at Facebook