Website access: separating DNS, CDN and origin failures

A browser needs to find the domain's destination, establish a connection and secure session, then receive an application response. Success at one stage does not prove the next stage works. Recording the actual failure is more useful than grouping every symptom under a generic unavailable label.

DNS connects domain names with information used to find their destinations. Cached answers mean that a record change does not necessarily reach every user at the same moment. Changing a resolver without evidence also cannot repair every kind of connection problem or server failure.

A CDN-backed website may receive a user's request at an edge service before contacting its origin when needed. Trouble on either part can produce different symptoms. A failed request to one domain is not sufficient evidence that every website using the same provider is unavailable.

Access can also differ between connections or regions. Testing from several points helps establish where the difference was observed. It does not independently establish whether the cause was destination policy, routing, DNS or another dependency. That conclusion requires evidence from the relevant stage.

Record the exact hostname, time, origin and error type. If the server returns an HTTP response, keep its status code too. Separate a connection timeout from a certificate error or an application failure. A simultaneous comparison with an independent domain helps establish how broadly the problem is shared.

Covered's Cloudflare access view reports tests of selected domains. Their selection is about connectivity testing, not popularity, endorsement or service quality. Provider traffic charts also describe the provider's dataset; they are not direct measurements of every internet user's traffic in Iran.