Why this matters
A website that “will not load” can fail at several different layers. The browser may be unable to find the server’s IP address, the server may present an invalid certificate, the connection may be too slow or interrupted, or the web server may accept the request and then fail while producing the page. These failures can look similar in a browser, especially when all the user sees is a spinner, a blank page, or a generic error message, but they require different remedies.
DNS provides the first useful split. A hostname such as www.example.com is a name, while a browser ultimately needs an IP address to establish a network connection. Once that connection is possible, HTTPS depends on a valid certificate for the requested hostname. A certificate may be expired, issued for a different name, or otherwise unacceptable to the client. After DNS and TLS are working, network performance and server behavior become more plausible causes. The tools in this guide help separate those possibilities in a repeatable order and produce evidence you can include when escalating the problem.
The sequence matters. Testing a generic speed result before checking DNS or the certificate can waste time, while changing server configuration before confirming whether name resolution works can hide the original cause. Record what happened at each stage, including the exact time and the exact error text. UTC time is useful here because local clocks can differ between the person reporting the problem and the person investigating it.
The process
-
Record the failure precisely and note the UTC time.
Open utc-time and record the UTC time shown by the tool. At the same time, write down the complete URL, the hostname, the approximate time of the failed attempt, the browser and operating system, and the exact message displayed. If the page remains blank, state that explicitly rather than describing it as a timeout. This evidence distinguishes a single failed request from a site that consistently fails and helps determine whether a later retest happened under comparable conditions.
-
Test whether the hostname resolves in DNS.
In dns-lookup, enter the hostname in Domain Name. Begin with Record Type set to A to look for an IPv4 address. Repeat with AAAA to check for an IPv6 address, and use CNAME when the site is known to be an alias rather than the final host. These checks answer a fundamental question: can the public DNS system map the requested name to the addresses it is supposed to use?
-
Interpret the DNS result according to the record type.
Do not treat every record type as part of the same web-hosting check. MX concerns mail delivery, TXT can contain verification or policy data, and NS identifies authoritative nameservers. PTR is used for reverse lookups, while SOA contains nameserver information. For a website load failure, A, AAAA, and CNAME are the most directly relevant. Save the returned records and note whether the hostname produced an address, produced only an alias, or returned no usable answer. A failed query or a “DNS query failed” message is evidence to retain, but it does not by itself identify whether the fault lies in the domain configuration, the resolver, or the network path.
-
Check the HTTPS identity separately from DNS.
Open ssl-checker, enter the same hostname in Hostname, and perform the check. This isolates certificate and TLS-related evidence from name resolution. Review whether the hostname is covered by the presented certificate and whether the certificate presents other problems reported by the checker. A certificate error can prevent an HTTPS page from loading even when DNS is correct. If the tool reports “Rate limit exceeded. Please wait a moment and try again,” record that response and retry later rather than treating the rate-limit message as proof that the site’s certificate is invalid.
-
Check whether the connection is being reported as failed at any layer.
Compare the DNS and certificate evidence before moving on. If the hostname does not resolve, investigate DNS first. If it resolves but the hostname check reports a certificate problem, investigate the certificate and its coverage of the requested name. If both checks are healthy, continue to network and server-side investigation. This order prevents a downstream symptom, such as a timeout, from being mistaken for the underlying problem.
-
Measure the network path independently of the website.
Open speed-test and run its download, ping, and upload measurements. The listed size choices are 5 MB and 50 MB, and the tool can produce a downloadable result. Use the measurement to determine whether the connection is broadly operational, unusually slow, or showing signs of interruption. If the tool reports “Failed to measure download speed,” “Failed to measure ping,” or “Failed to measure upload speed,” record which measurements failed. A failed generic test does not prove that the website itself is broken; it may indicate a local connectivity problem, a test endpoint problem, or an unstable path.
-
Retest the same evidence under the same conditions.
Repeat the DNS query and hostname check, and note the UTC time again. If possible, make the comparison from the affected network and from a materially different connection, such as another location or network type. Keep the hostname, record type, and certificate target constant. If the site works from one path but not another, the evidence points toward a network-path, DNS-resolution-path, or service-distribution issue rather than a universally broken deployment.
-
Escalate with a compact evidence bundle.
Provide the exact URL and hostname, the first and last UTC timestamps of testing, the DNS records returned for A, AAAA, and CNAME where relevant, the certificate-check result, and the speed-test measurements with any failure messages. Also include whether the failure affects every visitor or only some, whether it affects a hostname or a specific path, and whether HTTP and HTTPS behave differently. If DNS resolves, the certificate is acceptable, and the network test succeeds, but the page still fails, the remaining likely layer is the web server, application, redirect chain, or origin configuration. That is the point to involve the hosting or application owner with the collected evidence.
Worked example
Suppose store.example.com does not open for a customer. The customer sees “This site can’t be reached” and cannot tell whether the problem is the domain, the certificate, or the application. The support technician first opens utc-time and records 14:32 UTC. The customer’s report includes the full URL, the hostname store.example.com, the browser, and a screenshot of the error. The technician then opens dns-lookup, enters store.example.com in Domain Name, and queries A. The result contains an IPv4 address, so basic IPv4 resolution is working. An AAAA lookup returns no IPv6 address, which is useful to document but is not automatically an error. A CNAME query shows that the name is an alias, and the evidence confirms the expected DNS chain.
The technician next uses ssl-checker, enters store.example.com in Hostname, and finds that the certificate does not cover that hostname. The diagnosis can now be narrower: the name resolves, but HTTPS is presenting a certificate for a different identity. Changing application code or repeatedly retesting download speed would not address that certificate mismatch. The escalation to the hosting team includes the UTC timestamp, DNS answers, the exact hostname tested, and the certificate result. The hosting team can correct certificate deployment or the hostname that should be used.
In a contrasting case, suppose DNS returns 198.51.100.24 and the certificate check reports no issue. A speed-test then shows working ping, download, and upload measurements at the available 5 MB test size, while store.example.com still times out. The remaining evidence no longer points to basic resolution, certificate identity, or a generally unusable connection. The support ticket can state that the page still fails despite successful DNS, TLS, and network checks, and include the timestamps and returned records. That gives the application or origin owner a concrete starting point for investigating redirects, routing, origin availability, and server-side response generation.
Common mistakes
-
Testing only the root domain when the failure involves another hostname.
Services can use names such as www, shop, or an API subdomain with different DNS and certificate coverage. A healthy root domain does not prove that the failing hostname is configured correctly. Test the hostname copied from the failing URL.
-
Using a mail record as evidence that the website is configured correctly.
An MX lookup concerns mail routing. It may be relevant to a separate mail incident, but it does not establish that the web hostname has an A or AAAA address.
-
Changing the DNS record before preserving the original result.
Overwriting a record can erase the evidence that identifies the incident. Copy or record the returned values, the record type, and the query time before making configuration changes.
-
Interpreting a successful speed test as proof that the website works.
A speed test exercises the connection to its test endpoints. The site may still fail because of a redirect loop, a broken origin, a missing route, a server-side exception, or a hostname-specific configuration error.
-
Interpreting a failed speed test as proof that the site is down.
The failure may concern the test endpoint, the local connection, or network instability. It does not identify the website’s DNS, certificate, or application state.
-
Ignoring the requested hostname during the certificate check.
A certificate can be valid for one name and unusable for another. The check must use the same hostname the browser requests, not merely a similarly spelled domain or the organization’s main site.
-
Reporting only “the site is down.”
That statement does not say whether users receive a DNS error, a certificate warning, a timeout, or a blank response. Each symptom points toward a different investigation stage, and the exact message is essential.
Limitations of this approach
-
These checks do not reproduce every browser-specific failure.
DNS, certificate, and network behavior can vary by browser, operating system, network, DNS resolver, and security software. A successful check from one location does not guarantee success for every visitor.
-
DNS results are observations, not proof of application health.
A correct address merely directs the client toward a host. The server behind that address can still be unavailable or return broken content.
-
A certificate check does not validate the entire page.
It focuses on the HTTPS identity presented for the entered hostname. It does not prove that redirects work, scripts load, assets exist, or application logic completes.
-
Speed-test results are not a website performance analysis.
They measure the tool’s download, ping, and upload tests, not individual page resources, server response composition, browser rendering, or third-party dependencies.
-
Browser-based tools are the wrong instrument for definitive origin diagnosis.
When the basic checks pass, the necessary evidence may require access to origin logs, server configuration, routing, application monitoring, and the exact response from the failing request. Escalate to the relevant owner rather than continuing to infer the cause from external observations alone.
-
Public DNS and certificate observations can change.
Records may be updated, and certificate deployments may change while an incident is in progress. Timestamps and repeated checks therefore matter, especially when the first result and the escalation packet must describe the same incident window.
Frequently asked questions
Which check should I run first?
Check DNS first. If the browser cannot obtain an IP address for the hostname, the request cannot progress normally to HTTPS or the web server. Then check the certificate and measure the network.
What does a successful DNS lookup prove?
It shows that the queried resolver returned the requested DNS information. For a web hostname, an A or AAAA result provides an address that can be used for a connection. It does not prove that the destination is healthy or that the page will load.
Do I need both A and AAAA checks?
Both are useful when gathering evidence. A concerns IPv4 and AAAA concerns IPv6. The absence of an IPv6 result is not automatically a failure, particularly if the environment being investigated is IPv4-only.
Why check the hostname instead of the full URL?
DNS and certificates are associated with the hostname. A full URL can include a path, query string, port, or scheme, but those parts are checked later during connection and request handling.
Can a healthy certificate still cause a page not to load?
Yes. A valid certificate removes one class of HTTPS failure, but the server can still return an error, redirect incorrectly, omit required assets, or run application code that fails.
Why record UTC time instead of only local time?
Different systems and locations may use different local time zones. UTC provides a shared reference for correlating the browser report, DNS checks, certificate observations, network tests, and server logs.
What if the speed test cannot complete?
Record the specific failure message and keep the DNS and certificate results. An incomplete speed test leaves the network condition uncertain; it does not justify declaring the site down without checking the remaining layers.
When is the evidence sufficient to escalate?
Escalate when the failure is reproducible and the record includes the exact hostname, URL, UTC timestamps, browser details, DNS results, certificate result, and network measurements. If DNS, TLS, and connectivity all appear healthy but the site still fails, the evidence is especially useful to the hosting or application owner, who can investigate the origin and server-side request path.