Microsoft engineer on cloud outages: rigor and redundancy feed a failure loop
tianyin_xu · x · 2026-11-17
A Microsoft engineer comments on cloud outage discourse: moving to the cloud was supposed to solve this, but every incident triggers the same two predictable reactions — calls for more rigorous specification/formal verification/testing, and calls for more redundancy. Both add complexity, creating new failure modes that eventually require more of both. Lather, rinse, repeat.
More from Infra
- Beyond solar capture: the case for 'power liquidity' at point of use — edgarpavlovsky · 2026-11-17
- Drogue Raises Pre-Seed From Garuda VC to Power AI Data Centers With Ocean Thermal Energy — arian_ghashghai · 2026-11-17
- Cloud Cost Rule of Thumb: $1,000/Month Equals About $0.11/Hour — StewartalsopIII · 2026-11-17
- Analyst: Data Center Builder $BE Can Scale Operations Faster Than Expected — BenBajarin · 2026-11-17
- NVIDIA unveils Vera CPU rack-scale HPC platform with 2x x86 energy efficiency — luisceze · 2026-11-17
- AI Now calls BAAJA bill a "backdoor bailout" of AI industry ahead of Senate vote — AINowInstitute · 2026-11-17