Six Secrets to Succeeding with Microservices
summary
Microservices succeed only when architecturally significant constraints are chosen and supported by organization, teams, and automation rather than topology alone; Gartner predicts 90% fail prediction and many migrations take 3–5 years. Real numbers used as anchors: Amazon does 136,000 deployments a day, Shopify runs ~40 deployments/day and 1.3M requests/sec. Key practices called out: domain-driven boundaries, isolated databases, API-first design, testable governance with tools like ArchUnit, and evolving via a modular monolith to avoid distributed-monolith traps.
AI-written summary · about
chapters (8)
- 0:00Introduction and context
- 15:56Nobody wants microservices
- 23:25No single definition of microservices
- 36:30Trade-offs are more extreme than you think
- 45:25Decomposing monoliths is not always hard
- 50:14Granularity need not be absolute
- 54:48Monolith first is often practical
- 58:01Six secrets recap
related talks
23mMicroservices – Diptanu Choudhury
49mChallenges in Implementing Microservices
45mMicroservices: A retrospective
26mEnhance System Resilience and Efficiency with Microservices
50mChallenges in Implementing Microservices
47mMicroservices? Monoliths? An Annoying Discussion!
26mCreating a scalable ecosystem of Microservices
39m10 Tips for failing at microservices