If your website works for you but customers say it will not load, the problem may affect only certain networks, devices or routes to your hosting provider. A website down for some users needs a different response from a complete outage: first establish who is affected, then identify the failing connection before changing live settings.
Confirm whether the problem affects everyone or only certain visitors
Ask two people on different networks to open the exact same URL. One can use office Wi-Fi while the other uses mobile data. Test the main domain and its www version separately, and record the time, location at the broad network level, and exact error shown. An external availability check adds another perspective, but one successful test does not prove that every visitor can reach the site.
If a private browsing window works while a normal window fails, investigate cached redirects, cookies and extensions. If the site works over mobile data but not office Wi-Fi, investigate that network's DNS resolver, filtering and cached records. Do not ask customers to share passwords or other sensitive information.
Match the visible symptom to the likely cause
“Server not found” or DNS errors
DNS translates a domain name into the address needed to reach a service. A missing or incorrect record, outdated nameserver delegation or a cached negative response can make the same domain resolve differently for different visitors. Compare your domain and www records with the values expected by your hosting provider. Check where the authoritative nameservers are managed, especially after a domain transfer or hosting migration.
Cloudflare's DNS troubleshooting documentation explains why different resolvers can continue returning different answers after a change. DNS cache clearing on one device is not a universal fix; avoid changing records repeatedly without confirming which entry is wrong.
Certificate or secure-connection warning
A certificate warning means the browser could not establish the expected trusted HTTPS connection. Check whether the certificate is valid for both the main domain and www, whether renewal completed, and whether your hosting or CDN is presenting the correct certificate. Never tell visitors to bypass a certificate warning or enter information on an untrusted connection.
Timeouts and gateway errors
A timeout can indicate network routing, firewall restrictions or an overloaded server. A 502 or 504 response may point to a proxy that cannot get a usable response from the application behind it. Compare the affected URL against your hosting and CDN status pages, server logs and recent deployment history. A homepage can appear healthy from cache while a checkout or contact form fails, so test the actual affected feature.
Only one page or customer journey fails
If the homepage loads but a product page, booking flow or login does not, investigate that route and its dependencies. Check for broken redirects, access-control rules, recent application updates and failed API requests. Keep customer data out of screenshots and support tickets.
Use a safe troubleshooting order before changing production
- Capture the symptom: note the exact URL, error, approximate time and which networks are affected.
- Check external availability: test from more than one network and review provider incident notices.
- Verify domain routing: confirm registration status, nameservers and the expected DNS records.
- Verify HTTPS: check certificate coverage, expiry and CDN-to-origin configuration.
- Review the server and application: inspect response codes, error logs, resource pressure and recent changes.
- Make one controlled correction: document the original setting, change only the confirmed fault and have a rollback plan.
- Retest affected journeys: verify from the networks that originally failed, not only your own computer.
If you cannot identify the cause confidently, pause configuration changes and send your hosting or development team the captured evidence. Changing DNS, security rules and application settings simultaneously makes the root cause harder to find.
Consider search visibility and customer communication
Temporary access problems can affect customers and search crawlers. Google's HTTP status code guidance explains that persistent server errors can reduce crawling and eventually affect indexed pages. This is a reason to resolve availability issues promptly, not a prediction of an immediate ranking loss.
When an incident is confirmed, use an independent communication channel to acknowledge it. Explain which services are affected and when you will provide another update. Avoid promising a resolution time you cannot verify.
Reduce the chance of another partial outage
Monitor the main website and at least one important customer action, such as a contact form or checkout. Set alerts for certificate renewal and domain expiry, document DNS ownership, and keep a record of hosting and deployment changes. After each incident, record what failed, how it was fixed and how the same problem could be detected earlier.
If you need help tracing an intermittent outage or organizing preventive checks, DesDevSol's website maintenance and security services can provide a starting point for a focused review.