Revolut postmortem: 5-hour global cards outage caused by one GCP zone, no DB monitoring
cyb3rops · x · 2026-11-15
Revolut published the postmortem of its October 20 global cards outage that lasted 5 hours, after Gergely Orosz pressed publicly for three weeks.
Key findings:
- A power loss in one zone (Zone B) of Google Cloud's London region caused some databases to fail zonal failover, taking down global processing.
- Some quorum groups (5 coordination servers, 3 needed) already had one server offline without any monitoring alert; losing two more in Zone B left only 2 nodes—below quorum.
- Engineer tooling for monitoring and recovery also depended on Zone B, delaying remediation further.
Orosz calls the design alarming: a single-zone outage cascaded globally, suggesting weak disaster-recovery and failover processes.
More from Infra
- Alphabet's free cash flow turns negative for first time in 20 years, burning ~$15B over two quarters — luisdans · 2026-11-15
- Reddit debate: Do US chip export controls slow down global open-source AI software too? — No-Birthday-7360 · 2026-11-15
- Inference benchmark gotchas: bs=1-only engines are 'model runners,' warn practitioners — VoidAsuka · 2026-11-15
- China's state broadcaster claims all new electricity demand will be met by clean energy by 2030 — Polymarket · 2026-11-15
- ChatGPT verdict: NVIDIA's real product is market confidence, and a Lehman moment may be brewing — bingxu_ · 2026-11-15
- NVIDIA's 10-Q reveals billions in private-equity bets on KKR and Helix, drawing scrutiny — SumitGup · 2026-11-15