Do these 3 things before closing this tab:
1Repair Windows errors before they cause bigger problems2Fix the driver behind crashes, sound loss and screen glitches3Clear out junk files and repair common Windows errorsA broken website can fail at several different points: its domain may not resolve, the server may be unreachable, the application may return an error, or the page may load while its features fail. Start by recording the exact failure and checking its scope; then use DNS results, HTTP responses, and browser or server evidence to identify the layer that needs attention.
1. Establish exactly what is broken
Before changing settings, reproduce the problem and write down what happens. Note the complete URL, the time, the visible error, and whether the issue affects the whole site or only a page or feature. A useful record lets you compare tests and gives your hosting, DNS, or development team something specific to investigate.
- Try the URL in another browser and on another device.
- Test on a different network, such as a phone’s mobile connection rather than the same Wi-Fi.
- Check both
httpandhttps, and bothwww.example.comandexample.com, if those versions apply to your domain. - See whether every page fails or only one route, image, script, or other feature.
If the site fails only on one device or network, cached data, local DNS, or the network provider’s route may be involved. If it fails for multiple visitors across different networks, investigate DNS, hosting, a CDN, or the application itself.
2. Check whether DNS and the network can reach the site
If the browser says it cannot find the server, cannot resolve the hostname, or times out before showing an HTTP response, start with DNS and connectivity. Google notes that DNS errors, network-unreachable conditions, timeouts, connection refusals, and failed connections can prevent a server from returning any HTTP status at all. In that case, inspecting the page’s source will not explain the failure.
Verify the domain’s DNS configuration
- At your domain registrar, confirm that the domain uses the intended nameservers.
- Ask your DNS or hosting provider to confirm the authoritative records, especially the domain’s
A,AAAA, andCNAMErecords. - Compare DNS answers from more than one public resolver. Different answers can help identify stale or inconsistent DNS data.
- Ask the provider to confirm that DNS and hosting services are operating normally.
A missing or incorrect record can stop visitors from reaching the correct server. Avoid changing records at random: first establish which nameservers are authoritative and what values your host expects.
3. Read the HTTP status and redirect chain
If the hostname resolves and the server responds, record the HTTP status code and each redirect in the chain. You can do this in browser developer tools or with an HTTP header checker. Google explains that status codes are generated by the hosting server; the code is a clue about what happened, not a complete diagnosis.
| Result | What it can indicate | What to investigate |
|---|---|---|
| 4xx response | A client-side, authorization, or missing-resource problem. | Check the requested URL, access permissions, authentication rules, and whether the resource exists. |
| 5xx response | A server-side failure or capacity problem. | Check hosting status, server and application logs, recent changes, and available capacity. |
| Redirect loop or malformed redirect | A redirect or site-configuration problem. | Review redirect rules and confirm that HTTP/HTTPS and www/non-www versions lead to the intended canonical URL. |
| Success response, but the page is broken | The server returned a successful status even though the content or its features may be faulty. | Inspect the response body, browser console, and failed resource requests; continue to step 4. |
A normal page should generally return a success response. When a page has intentionally moved, it should use the appropriate permanent or temporary redirect rather than an improvised chain or loop.
4. Look for application and front-end failures
An HTTP 200 response does not prove that visitors received a working page. Google describes an error-like page that returns 200 as a “soft 404.” A broken database connection or a missing JavaScript file can also leave a page unusable despite a successful status.
Check the browser and server evidence
- Open the browser’s developer tools and look for failed JavaScript, CSS, image, or API requests and errors in the console.
- Review web-server and application logs around the time the failure occurred.
- Check database health and the environment variables or configuration the application needs.
- Review recent deployments or configuration changes. If a recent change is the likely cause, consider a controlled rollback.
- Check CDN or cache behavior, firewall rules, and bot protections that might block requests or serve stale content.
Use the evidence to narrow the responsible layer. For example, a failed script request points to a different issue than an application log showing database errors. Preserve the exact URL and timestamp while reproducing the fault so the relevant team can match it to logs.
5. Confirm recovery and check Google’s view
Once you have corrected the underlying problem, test the affected URL again from more than one network and verify that its response, redirects, and page features now work. If Google indexing matters for the page, use URL Inspection in Search Console and the Crawl Stats report to see whether Googlebot can fetch it and whether crawl failures cluster around DNS, connectivity, robots.txt, timeouts, or HTTP errors.
Google’s current Crawl Stats documentation sets a DNS-resolution failure rate above 5% of requests in a day as an issue threshold in the report’s host-status view. This is a Googlebot-specific diagnostic threshold, not a general uptime target. After fixing the cause, request inspection or recrawling where appropriate and keep monitoring server logs.
Google warns that returning 503 or 429 responses continuously for more than about two days during an overcrawling emergency can cause affected URLs to be dropped from its index. Those codes may be appropriate in some temporary situations, but do not leave them in place longer than necessary; verify recovery and monitor the affected pages.
The Tool Desk
Outbyte PC Repair FREERepair Windows errors before they cause bigger problemsFix Now →Outbyte Driver Updater FREEFix the driver behind crashes, sound loss and screen glitchesFind Drivers →Choose the next action by the evidence
When deciding what to change or whom to contact, match the response to where the failure occurs, who is affected, and how reversible the fix is. DNS or network failures call for DNS-provider, host, or connectivity evidence; 4xx, 5xx, and redirect failures point toward different server or configuration checks; a broken page with a 200 response needs application or front-end investigation. Consider whether a quick rollback is possible and whether the fault also threatens Google’s ability to crawl the page. Search Console shows Googlebot-specific results; an external uptime check can provide ongoing checks from outside your network.
For continuous monitoring, Google Cloud Monitoring uptime checks can verify status codes and, when configured, response content. By default, HTTP uptime checks verify that the response code is 2xx, so a check based only on status may not catch an error page that incorrectly returns 200.
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
