Insights:What to Check First When Domain or DNS Issues Occur: A 5-Step Troubleshooting Guide
First confirm the symptom; don't immediately edit records
| Symptom | Likely cause |
|---|---|
| Unreachable from all regions | Authoritative DNS, domain status, origin server |
| Only some regions unreachable | Local cache, routing, CDN nodes |
| Intermittent availability | TTL, load balancing, DNSSEC, resolution routing |
| Site works but email bounces | MX, SPF, DKIM, DMARC |
| Domain just renewed or migrated | Registrar status, NS switch, record sync |
After recording "who, where, when, which address was accessed, and what error appeared," you can proceed with each layer of analysis without guessing.
Step 1: Check DNS records
Query A, AAAA, CNAME, and MX from both local and public resolvers:
dig +short A example.com
dig +short AAAA example.com
dig +short CNAME www.example.com
dig +short MX example.com
Key points to check:
- Whether returned IPs match the current origin server or CDN.
- Whether CNAME points to a decommissioned old domain.
- Whether MX records route to active mail services.
- Consistency across different resolvers.
Step 2: Check authoritative servers
dig +short NS example.com
dig +short SOA example.com
dig +trace example.com
Check whether the NS records belong to your current DNS provider, and whether the SOA serial is updating. If the NS still points to the old provider, no amount of changes in the admin panel will take effect.
Common issues:
- The domain was just migrated, and the NS at the registrar hasn't been updated yet.
- One of the authoritative servers is unreachable, causing intermittent resolution failures.
- Resolution records exist at both the old and new providers and are inconsistent.
Step 3: Rule Out Cache and TTL
DNS records have a TTL. After a change, old records may still persist locally, in the OS, routers, ISPs, and CDNs.
When troubleshooting, query at least three resolvers:
dig @8.8.8.8 +short A example.com
dig @1.1.1.1 +short A example.com
dig @223.5.5.5 +short A example.com
If public resolvers already return the new address but your computer still shows the old one, the problem is local cache, not the authoritative server. Don't keep changing records at this point.
Step 4: Check Links and Regional Differences
The same domain can resolve differently across regions, commonly due to:
- Abnormal CDN routing.
- Poisoned cache or misconfigured intelligent routing sending traffic to the wrong path.
- A faulty link between an ISP and the origin server.
- A firewall or WAF blocking by region.
Use online DNS testing tools from multiple regions to compare returned IPs and response times. If necessary, access the origin IP directly to determine whether it's a "DNS resolution error" or an "unreachable origin."
Step 5: Check the Application Layer
A correct DNS resolution doesn't guarantee the site works. Continue investigating:
- Whether the HTTPS certificate has expired or the domain doesn't match.
- Whether HTTP status codes are 502, 503, or 403.
- Whether origin server ports, firewalls, and load balancers are functioning normally.
- Whether the email SPF, DKIM, and DMARC records are properly configured.
curl -I https://example.com/
curl -o /dev/null -s -w "ttfb=%{time_starttransfer} total=%{time_total}\n" https://example.com/
Maintain a Troubleshooting Log
We recommend recording each instance:
| Field | Example |
|---|---|
| Time | 2026-11-02 17:15 |
| Source | Local machine / Guangdong / Overseas |
| Query Type | A, MX, CNAME |
| DNS Resolver | 8.8.8.8 |
| Return Value | 1.2.3.4 |
| TTL | 600 |
| Conclusion | Authoritative DNS record has been updated, but local cache hasn't expired |
This table helps avoid disputes like “I thought we changed it yesterday,” and enables quick comparisons during future incidents.
Things to Avoid
- Don’t delete DNS records without first creating a backup.
- Don’t temporarily lower TTL values during an incident—first assess the impact scope.
- Don’t draw conclusions based on a single DNS resolver.
- Don’t mix DNS issues with HTTPS, origin server, or mail authentication problems when making changes.
The core of DNS troubleshooting is not “tweak until it works,” but identifying which layer isn’t performing as expected. Each layer has evidence, so fixes won’t keep coming back.
Related reading: