On 12 June 2025 Google Cloud suffered a global outage caused by a faulty automated policy update in its service control layer. Google's own products were affected, and so were a long list of third-party services that build on Google Cloud — including tools that most people would never associate with Google.
The pattern by now should be familiar
Three of the biggest outages of the last two years — AWS, Cloudflare, Google — all followed the same shape: an internal change, automatically distributed, that broke a control system everything else depended on. None of them were attacks.
Where that leaves a small business
The honest answer is that you cannot prevent this and shouldn't spend money trying. What's worth doing is cheap:
- Map your dependencies once. Which tools do you genuinely need to trade today? Usually phones, email, payments and one core application.
- Make sure those four aren't all on the same platform. If they are, change one of them.
- Have a manual mode. Paper order pad, phone list, card machine that works on mobile data.
- Communicate. A short message to customers during an outage prevents most of the damage to your reputation.
We track vendor status feeds for the services our customers rely on and publish here when something is widespread.