What is a CDN, and why do outages spread so widely?

A content delivery network, or CDN, is a group of servers spread around the world that sit between visitors and a website. CDNs make sites faster and tougher against attacks. They also explain why one company’s fault can suddenly take down thousands of unrelated sites.

The problem a CDN solves

Distance matters on the internet. If a site’s only server is in one country, a visitor on the other side of the world waits for every message to cross that distance. A CDN keeps copies of the site’s files in many cities. A visitor is sent to a nearby copy, so images, scripts and often whole pages arrive far sooner.

What a CDN does besides speed

• Absorbs traffic spikes. Millions of requests are spread across many servers instead of landing on one.
• Filters attacks. Flood attacks are soaked up at the edge, and many CDNs also block known bad traffic and automated abuse.
• Handles encryption. The CDN can hold the site’s certificate and manage HTTPS on its behalf.
• Keeps a site up when the origin struggles. If the real server is slow or briefly unavailable, the CDN can sometimes keep serving saved copies.

The cost: concentration

A handful of providers carry a very large share of the web’s traffic. That is efficient, but it means many unrelated sites share the same front door. When that front door has a problem, they all show errors at once, even though their own servers are fine. From a visitor’s point of view it looks as though half the internet has broken.

This is a recurring pattern in major outages. The fault is often in a shared service, such as a content delivery network, a cloud region, a DNS provider or an identity service, rather than in any one website. Our guide to major internet outages gives real examples.

What a CDN error looks like

When a CDN cannot reach a site’s own server, or has an internal fault, visitors see an error page that often carries the provider’s name. Typical signs:

• A 5xx error such as 502, 503 or 504, sometimes with a branded message. See what the 5xx errors mean.
• A page saying the site cannot be reached because of an error with the network in front of it.
• A “checking your browser” or challenge page that never finishes, which is a security filter not a server failure.

Why we sometimes say “blocked”

CDN security filters can also refuse automated checks. When a site answers our probes with a challenge instead of the page, we cannot see what a real visitor sees. We show that as Blocked and explain that it does not mean the site is down. Why some sites block automated checks explains this in more detail.

Working out what is affected

1. Check more than one site. If several unrelated sites fail in the same way at the same time, suspect a shared provider.
2. Look at the provider’s status page. Large providers publish incident updates, and these are usually the fastest reliable source.
3. Check our outage map. Clusters of reports across different services are a strong sign of a shared cause.
4. Wait, then retry. Most large provider incidents are fixed within hours, and recovery can be uneven while caches and connections settle.

For site owners: reducing the risk

• Know exactly which providers your site depends on: DNS, CDN, hosting, certificates, email, payments and login.
• Make sure your status page and alerts do not depend on the same provider as your main site, or they will fail together.
• Consider whether a second provider for DNS is worth the extra complexity for your site. It protects against one specific kind of failure.
• Prepare a simple message you can post in minutes saying you are aware and working on it.
• Monitor from outside your own network. Our monitoring guide shows how.

The takeaway

CDNs are a big reason the web feels fast and survives attacks. The trade-off is that a few shared services sit under a huge number of sites. When many unrelated sites fail together, the cause is usually one of those shared layers, and the practical advice is to confirm it, check the provider’s status page and wait for them to fix it.

Related guides

• Why big websites go down: cloud, CDN and DNS failures
• Major internet outages and what caused them
• How websites stay up: redundancy, failover and backups

Last updated 2 November 2026. Written by the siteupdown.com team. Spotted a mistake?

← All help guides