Question 1 of 6

Site down, your name servers are down

Customer ticket

Subject: Site down, your name servers are down

I am having trouble reaching my website http://example.com/ from different networks. Are your nameservers working properly?

2. Thought process

The customer has already decided it’s “your nameservers”. My job is to reassure them with evidence, not just deny it, and then lead them to the real cause. I split the problem into two layers and work through them in order:

  1. Platform first. I check cloudflarestatus.com and internal incident channels, and query the zone’s assigned nameservers directly (dig @xxx.ns.cloudflare.com example.com). If they answer authoritatively, the nameservers are fine and I can say so with confidence.
  2. Delegation. I compare the NS records at the TLD (dig +trace / whois) with the pair assigned in the dashboard. A mismatch, leftover old nameservers, or a zone that is Pending or Moved is a classic cause of “down on some networks”, because different resolvers cache or choose different nameservers.
  3. DNSSEC. “Different networks” is a strong hint here. A stale DS record breaks resolution only on validating resolvers, which is exactly the inconsistent behaviour the customer describes. Comparing a query with and without +cd confirms it quickly.
  4. Records. I check that the apex and www records exist, and whether they are proxied or DNS-only.
  5. HTTP layer. If DNS resolves, “site down” is really an origin or edge question. The 521, 522 and 523 errors and the Ray ID tell me whether Cloudflare can reach the origin. I’d check the zone’s Analytics for a spike in 5xx errors.

The customer used example.com. In a real ticket that would be their actual domain, but example.com itself is reserved by IANA (RFC 2606) and is not a real Cloudflare zone, so I’d check that the right hostname was given. I ask for the error text, the Ray ID and command output from an affected network because the issue is network-dependent, and I need to see what their resolvers are returning.

3. Tools I would use, and why

ToolWhy
cloudflarestatus.comRule out a platform-wide or regional incident before blaming customer configuration.
dig / dig +trace / dig @<cf-ns>Check delegation from the root down, confirm that Cloudflare’s nameservers answer authoritatively, and compare results between resolvers (1.1.1.1, 8.8.8.8, the ISP resolver).
dig +dnssec / +cd, DNSVizConfirm or rule out a DNSSEC chain-of-trust break (stale DS record), the most common cause of inconsistent resolution.
whois / registrar lookupSee which nameservers and DS records the registrar has actually published.
whatsmydns.net / DNSCheckerSee how resolvers around the world answer, which matches “different networks”.
nslookupA familiar option for Windows users if they can’t run dig.
curl -svo /dev/null http://example.com/Once DNS works, see which IP was reached, the HTTP status, Cloudflare error codes (52x) and the cf-ray header.
Cloudflare dashboard (DNS, Overview, Analytics)Check zone status, assigned nameservers, records, proxy status, DNSSEC state and 5xx trends.