RunTheTests
/

Multi-Resolver DNS Propagation

Compare three public resolvers — conclusive when they disagree, honest when they agree.

The domain to look up

Queries go straight from your browser to public resolvers and registry services. Nothing about the domain you look up passes through this site.

How this check works

The same query goes to Cloudflare, Google and Quad9 over DNS-over-HTTPS, and the answers are compared. The asymmetry is the point. If they disagree, a change is definitely still working through, and that is a genuinely useful thing to know. If they agree, all you have learned is that these three resolvers hold the new record — the resolver your customer's ISP runs may still be serving the old one until the previous record's TTL expires, and nothing queried from here can see that. Tools that draw a world map are checking a longer list of the same kind of resolver.

What the results mean

Resolvers agree
These three hold the same records. A useful signal and not proof that every cache has updated.
Resolvers disagree
Conclusive: a change has not finished spreading. Wait for the old record's TTL to expire.
TTL
How long a resolver may keep serving an answer. It is the previous record's TTL that governs how long a change takes, not the new one's.
No answer
A resolver that did not respond. One quiet resolver is usually transient; all three is the domain.

Common problems and fixes

The change looks live here but not for my client
Their resolver is still inside the old record's TTL. Nothing speeds that up — lowering the TTL helps the next change, not this one.
It has been days and one resolver still disagrees
Check for two conflicting records at the authoritative nameservers, or an old record on a nameserver that is still delegated but no longer maintained.
I want to see this from other countries
Public resolvers are anycast, so a query from here may already be answered in another region. What a resolver comparison cannot do is speak for every ISP resolver, wherever it runs.

Frequently asked questions

How long does a DNS change take to take effect?

Up to the TTL of the record you replaced, counted from the last time each resolver cached it. If the old record had a TTL of 24 hours, some resolvers will serve it for another 24 hours no matter what you do now. Lowering the TTL before a change is what makes the change fast; lowering it afterwards helps the change after this one.

Can I check DNS propagation worldwide?

Not honestly, and no tool can. What is checkable is a list of public resolvers — this uses three. Disagreement among them proves a change is still spreading; agreement proves those resolvers are updated and nothing about the thousands of ISP resolvers that actually serve your visitors.

Why do two resolvers give different answers?

Either a change is still spreading, or the authoritative nameservers genuinely disagree — which happens when one nameserver in a delegated set was never updated. The second is worth ruling out if the disagreement outlasts the TTL.

More in WebAudit