The cloud is reliable most of the time, but when it fails, it fails at scale.

Recent examples

– AWS, October 20, 2025: a DNS automation problem in the US-East-1 region disrupted many internet services and apps.

– Microsoft Azure, October 29, 2025: a configuration change in Azure Front Door caused a global disruption of roughly eight hours, affecting Microsoft 365 and other services.

– Cloudflare, November 18 and December 5, 2025: two incidents disrupted a significant share of web traffic, including well-known sites and apps.

– Microsoft, July 23, 2026: a faulty network maintenance automation process in the West US region removed key network routes, disrupting Azure and Microsoft 365 services such as Outlook, Teams, and SharePoint for several hours, according to reports.

What these have in common

– Configuration and automation errors, not dramatic cyberattacks, were the cause in most cases

– Concentration: a small number of providers support a huge share of services, so one failure ripples widely

– Hidden dependencies: many organizations depend on providers they never contracted with directly

Lessons for your business

1. Know your dependencies. List the cloud services and platforms your business relies on, including the ones behind your software.

2. Plan for outages. Decide what staff do when email, chat, or a key application is down. Keep contact numbers and procedures accessible offline.

3. Have alternative communication channels, such as mobile phones and a separate messaging method.

4. Set recovery expectations. Match critical systems to realistic recovery goals (RTO and RPO).

5. Back up your cloud data independently. Outages are different from data loss, but a backup outside the provider gives options.

6. Monitor provider status pages and alerts, and subscribe to service health notices.

7. Review SLAs, and understand what compensation, if any, you receive. It rarely covers your actual losses.

8. Consider resilience where it matters most, such as redundant internet links and key systems with failover. Multi-cloud adds complexity, so use it selectively.

9. Test your continuity plan. Run a tabletop exercise for “Microsoft 365 is down for eight hours.”

Sources:
https://aws.amazon.com/message/101925/
https://www.reuters.com/technology/microsoft-azure-down-thousands-users-downdetector-shows-2025-10-29/
https://blog.cloudflare.com/18-november-2025-outage/
https://blog.incidenthub.cloud/azure-west-us-outage-jul-23-2026

DBR Soft can help you assess cloud dependencies and design practical continuity plans. Contact us for more information.