Cloud & CDN Status
One cloud and CDN provider is down.
Cloud and CDN outages are almost always regional and almost always confined to one service, which is why a global up or down verdict for a provider this size is close to meaningless on its own. A compute failure in one region can leave identical resources elsewhere completely healthy, and the provider's own dashboard is the only place with a genuine per-service, per-region picture.
What makes these incidents feel bigger than they are is concentration. A single region hosts the control planes for a great many products, so one regional fault produces symptoms in applications all over the world that have nothing obvious to do with each other. Edge networks amplify the same effect in the other direction: a configuration mistake propagates in seconds and takes thousands of unrelated sites down together, which is why the first sign is usually a pattern rather than one error.
We check each provider from a single vantage point, so a regional outage that does not touch our route will read as healthy here. Treat a green verdict as one data point and the provider's dashboard as the authority.
- up
- slow
- down
- no reading
Common questions about cloud and CDN providers
Check the CDN and DNS layers first, since those produce the widest blast radius. If unrelated sites are all failing at once, the cause is far more likely to be a shared edge network or DNS provider than any individual site. Reaching a site by IP while its hostname fails points squarely at DNS.
We check from one location and one endpoint. Cloud incidents are usually scoped to a region or a single service, so a fault that takes your instances offline can be entirely invisible from our vantage point. That is a real limit of an external check rather than a contradiction.
Test the same operation against a second region or a second network. A provider incident affects a whole region or service uniformly, while a security group, IAM policy, or DNS record change fails narrowly and consistently. If one region works and another does not, it is likely theirs. If both fail identically, look at what changed on your side.