1. Confirm the scope of the outage
Check whether the problem affects one page, the whole website, one network or every visitor. Test the hostname and, when appropriate, the server address separately. A scoped failure is easier to diagnose than the generic statement that “the site is down.”
2. Check the domain and DNS path
Confirm the domain has not expired and that authoritative DNS answers point to the intended service. Check A, AAAA and CNAME records and remember that cached answers can remain visible until their TTL expires.
3. Check the Cloud Server or hosting service
Verify that the server or hosting account is running and reachable. If the infrastructure resource is stopped, suspended or incorrectly addressed, there is little value debugging the application first.
4. Check the web stack and database
Confirm the web server, application runtime and database are healthy. Look for exhausted memory, full storage, failed services, connection errors and recent deployment changes.
5. Check TLS and the application itself
An expired certificate, redirect loop, failed dependency or application exception can make an otherwise healthy server appear unavailable. Use logs and the actual HTTP response to distinguish infrastructure failure from application failure.
Make the checks repeatable
Document the normal DNS target, server relationship and recovery steps for important sites. CloudBalkan AI Operator is built around this same idea: inspect live infrastructure context first, then propose an approved operation rather than guessing.