Major Outage Database
The most expensive downtime incidents in tech history. Real companies, real costs, real lessons.
| Company | Date | Duration | Est. Cost | Cause |
|---|---|---|---|---|
| CrowdStrike | Jul 2024 | Widespread, multi-day | $5.4B | Faulty content update caused Windows BSOD globally, grounding flights and disrupting hospitals |
| Southwest Airlines | Dec 2022 | Multiple days | $800M | Legacy scheduling system collapsed during winter storm, stranding millions of passengers |
| TSB Bank | Apr 2018 | Weeks | $370M | Failed IT migration from Lloyds platform left 1.9M customers locked out of accounts |
| AWS us-east-1 | Dec 2021 | ~5 hours | $150M | Network device overload cascaded across internal services, taking down major websites |
| British Airways | May 2017 | 3 days | $150M | Power supply failure at data center caused total IT system collapse, 75K passengers stranded |
| Delta Air Lines | Aug 2016 | ~5 hours | $150M | Power outage at operations center cascaded through systems, cancelling 2,300+ flights |
| Meta / Facebook | Oct 2021 | ~6 hours | $100M | BGP routing misconfiguration during maintenance made all Facebook services unreachable |
| Rogers Communications | Jul 2022 | ~15 hours | $100M+ | Routing filter deletion during maintenance cascaded, knocking out Canada-wide network |
| Roblox | Oct 2021 | 73 hours | $25M+ | Hash table performance issue in Consul service discovery under high load |
| Google Cloud | Jun 2022 | ~2 hours | $20M+ | Configuration change caused traffic blackholing across multiple regions |
| Fastly CDN | Jun 2021 | ~1 hour | $15M+ | Single customer config change triggered bug, taking down Amazon, Reddit, Gov.uk, and more |
| Cloudflare | Jun 2022 | ~2 hours | $10M+ | BGP change in 19 data centers caused widespread outage affecting millions of sites |
| Microsoft Azure | Jan 2023 | ~6 hours | $10M+ | WAN configuration change impacted connectivity between Azure regions globally |
| Slack | Feb 2022 | ~5 hours | $8M+ | Database infrastructure issue during configuration change disrupted all messaging |
| GitHub | Mar 2018 | 24 hours | $5M+ | Network partition caused database inconsistency requiring careful failover and data reconciliation |
Industry Benchmark
The average cost of IT downtime is $5,600 per minute according to Gartner research. For large enterprises, this figure can exceed $11,600 per minute.