Four major outages in the past year were caused by configuration changes and bugs, not attacks. Even small websites go down when everything depends on one provider.
Over the past year, three of the world's largest infrastructure providers — AWS, Microsoft Azure, and Cloudflare — suffered outages that left thousands of websites and apps unreachable for hours. Notably, the causes were not cyberattacks but configuration changes and software bugs.
Four major incidents
| Date | Provider | Cause | Impact |
|---|---|---|---|
| 19–20 Oct 2025 | AWS us-east-1 | A DynamoDB DNS automation bug left an empty DNS record. | Spread to EC2, Lambda, IAM, and other services for almost 15 hours. |
| 29 Oct 2025 | Azure Front Door | An inadvertent configuration change. | Azure and Microsoft 365 unreachable for about 8 hours. |
| 18 Nov 2025 | Cloudflare | A database permissions change made the Bot Management feature file balloon. | Many sites returned 5xx errors for about 6 hours. |
| 20 Feb 2026 | Cloudflare | A cleanup-task bug withdrew about 25% of BYOIP prefixes via BGP. | BYOIP customers' services unreachable for about 6 hours. |
Official reports: AWS, Cloudflare 18 November, Cloudflare 20 February, and a summary of the Azure incident from Dataconomy.
The pattern: small mistakes, big impact
All three providers run data centres in many countries, but their control systems are centralised. One wrong configuration change can spread worldwide within minutes. In the AWS incident, for example, a single empty DNS record in one region made many other services fail because they depend on each other.
Why small websites are affected too
Many small websites use one provider for everything: DNS, CDN, hosting, and email. When that provider has a problem, everything stops at once — and you cannot move anything because its control panel is down too.
Simple steps to reduce the risk
- Keep your DNS provider separate from your hosting/CDN, so you can still repoint your domain when hosting fails.
- Use a moderate TTL (e.g. 1 hour) so emergency changes spread quickly — see TTL and DNS propagation.
- Keep email separate from website hosting, so email keeps working when the site is down.
- Keep backups of your website and data in a different location, and write down how to move them.
- Follow your providers' status pages and have a way to update customers (e.g. social media or WhatsApp).
Sources
- Summary of the Amazon DynamoDB service disruption in us-east-1 — AWS — aws.amazon.com
- Cloudflare outage on November 18, 2025 — Cloudflare — blog.cloudflare.com
- Cloudflare outage on February 20, 2026 — Cloudflare — blog.cloudflare.com
- Azure outage: Microsoft blames Azure Front Door — Dataconomy — dataconomy.com
Written from the sources above as of 31 October 2026. Third-party policies and figures can change — check the original sources for the latest.
All news