When a domain still shows an old website, fails to resolve, or stops delivering email after a change, the cause is not always “DNS propagation.” This reusable checklist helps you separate caching from incorrect nameservers, record conflicts, hosting settings, and email authentication errors so you can test the right layer before changing anything else.
Overview
DNS translates a hostname such as www.example.com into information that browsers, mail systems, and other services can use. A domain’s registrar, authoritative DNS provider, recursive resolvers, hosting platform, proxy, and application may all be involved in a single connection. A problem at any one of these layers can look like a propagation delay.
Start by writing down the intended configuration. Record the domain, the affected hostname, the expected record type and value, the time of the change, and the service that should receive the traffic. This simple inventory prevents troubleshooting from becoming a sequence of guesses.
For example, a typical web setup might use an A record for the root domain, a CNAME record for www, and separate MX, TXT, or service-specific records for email and verification. Your hosting provider may require a different arrangement. Follow its exact hostname and value instructions, and do not assume that an example for one provider applies to another.
Before editing records, confirm which nameservers are authoritative. The DNS dashboard where you made the change is not necessarily the provider currently answering for the domain. A nameserver mismatch is one of the most common reasons a correct-looking change has no visible effect.
For a broader process, see the DNS propagation check guide and the walkthrough on using dig, nslookup, and whois.
Checklist by scenario
The domain or subdomain does not resolve
- Check the spelling, including the hostname and domain suffix.
app.example.com,www.example.com, andexample.comare separate names. - Confirm the domain is registered, has not expired, and uses the nameservers you expect.
- Query the authoritative nameserver directly rather than relying only on a public propagation checker. For example:
dig example.com NS, followed bydig @ns1.example-dns.com app.example.com. - Verify that the record exists at the authoritative provider with the correct type, name, and value. Some dashboards expect the host field to contain only
app, while others accept the full hostname. - Check whether a disabled, filtered, or proxied record is changing the response you expect.
If the authoritative response is wrong, this is a configuration problem rather than a resolver cache problem. Correct the record at the authoritative DNS provider, then test again.
The site resolves, but the wrong website appears
- Compare the returned IP address or canonical name with the destination supplied by your hosting platform.
- Look for duplicate or conflicting
A,AAAA, andCNAMErecords. A hostname generally should not use aCNAMEalongside other data at the same name. - Check both the root domain and
www. They may point to different services. - Confirm that the domain has been added to the hosting project, not merely configured in DNS.
- Review redirects, proxy settings, and SSL mode if the DNS result is correct but the browser reaches an unexpected destination.
For deployment-specific checks, use the guides for connecting a static site to a custom domain or running a Node.js app behind Nginx. If the browser reports a redirect loop after a proxy or SSL change, see the redirect troubleshooting guide.
Email stopped working after a DNS change
- Check the
MXrecords for the root domain. Their targets must match the mail provider’s instructions, and their priority values should be entered exactly as specified. - Confirm that an unrelated old mail provider is not still listed unless you intentionally use it.
- Review the provider’s required
TXTrecords for SPF, DKIM, and verification. - Ensure the SPF policy is published as one coherent record rather than several separate SPF records.
- Check that the DKIM hostname and value match the selector supplied by the mail provider.
- Publish a DMARC record only after you understand the reporting and enforcement settings you intend to use. Start with a carefully considered monitoring approach if you are still validating legitimate senders.
DNS can publish the correct email records while the mail service remains misconfigured. Check both sides: the public DNS answer and the sending provider’s domain setup. This is the practical core of an SPF, DKIM, and DMARC setup.
A change works for some people but not others
- Query several recursive resolvers and compare their answers.
- Test from a different network, such as mobile data, without treating one browser result as definitive.
- Clear local DNS cache only after recording the current result. This can help your device request fresh data, but it cannot change an incorrect authoritative record.
- Check the record’s TTL. A resolver may retain an older answer until its cached lifetime expires.
- Allow for dependent caches, browser behavior, and hosting-platform configuration to catch up after the DNS answer is correct.
Do not repeatedly edit the record while waiting. Every change can make the final state harder to identify and may introduce a new cached answer.
What to double-check
Nameservers and delegation
Use the registrar or a whois lookup to identify the delegated nameservers. Then query those authoritative servers directly. If your domain uses one provider’s nameservers but you edited another provider’s DNS zone, no amount of waiting will make that edit active.
Record names and formatting
Check for accidental duplication of the domain name. Depending on the dashboard, entering app.example.com in a field that automatically appends example.com can create an unintended name. Also inspect trailing dots, quotation marks, spaces, IPv6 records, and copied values. TXT records are especially easy to damage when a long value is wrapped or altered.
Authoritative versus cached answers
Compare a direct authoritative query with a normal recursive query:
dig @AUTHORITATIVE_NAMESERVER app.example.com A
dig app.example.com A
dig example.com MX
dig example.com TXTReplace the placeholder with a real authoritative nameserver and remove the extra leading space before the final command if copying it into a shell. The first query helps establish the source of truth; the later queries show what a resolver currently returns.
Hosting and application configuration
DNS only directs traffic. The destination must also accept the hostname. Confirm the domain is attached to the correct project, virtual host, load balancer, or reverse proxy. Confirm that the service listens on the expected port and that its certificate covers the hostname. If you are deploying a containerized application, the DNS record is only one step in the release path; the Docker, GitHub Actions, and Nginx deployment guide covers the surrounding workflow.
Common mistakes
- Changing records at the wrong provider: The visible dashboard is not proof that its nameservers are delegated.
- Using a CNAME at the root when the provider does not support it: Follow the provider’s documented root-domain option, which may use an alias-style record.
- Leaving old records in place: Stale A, AAAA, MX, or verification records can send different clients to different services.
- Confusing DNS with SSL: A correct DNS answer does not prove that HTTPS is configured or that the certificate matches.
- Assuming a propagation checker is authoritative: Check the authoritative response and several recursive resolvers.
- Publishing multiple SPF records: Consolidate authorized senders according to your mail provider’s instructions.
- Testing only one hostname: Root,
www, API, staging, and mail hostnames may have independent records. - Editing during an incident without a record of changes: Save the previous values and note each modification, time, and reason.
When to revisit
Return to this checklist whenever you change nameservers, move hosting, add a CDN or proxy, migrate email, rotate DKIM keys, change SPF senders, introduce a staging environment, or automate DNS through an API or Terraform. Recheck the workflow before seasonal planning cycles and whenever deployment tools or hosting providers change.
After a successful change, capture the final state: authoritative nameservers, important record names and types, hosting destinations, mail-provider requirements, TTL choices, and the commands or checks that confirmed the result. Remove obsolete records only after verifying that no active service depends on them.
For a practical final pass, work in this order: document the expected state, verify delegation, query the authoritative server, compare recursive results, confirm the hosting or mail service, then test from more than one network. If the authoritative answer is correct but clients still see an older value, wait for caching to expire rather than making another speculative edit. If the authoritative answer is wrong, fix the configuration at its source. This distinction is the fastest way to fix DNS not propagating without turning a small change into a larger outage.