Check whether a DNS change has reached the major public resolvers, and see where they still disagree.
When you change a DNS record, the new value does not appear everywhere at once. Resolvers around the world keep serving the answer they cached until that cache expires, so for a period some people see the new value and others still see the old one. This tool asks ten major public resolvers the same question at the same time and shows you who has caught up.
Enter a domain, pick a record type, and compare the answers.
It queries ten well-known public resolvers directly and compares their answers, flagging any that differ from the majority. That answers the question people usually have: has the change reached the resolvers my visitors actually use?
It is not a geographic propagation map. Our queries run from a single server in the United States, and the large public resolvers are anycast — a query to 8.8.8.8 is answered by whichever Google node is closest to us, not by one in Tokyo or São Paulo. Any tool that genuinely shows propagation by country has to run probes in each of those countries. We would rather tell you that than draw a map we cannot honestly fill in.
Almost always: caching. Each record carries a TTL telling resolvers how long they may reuse a cached answer. A resolver that fetched your record shortly before you changed it is entitled to keep serving the old value until that TTL expires, and it is not doing anything wrong.
So a mixed result right after a change is normal. What matters is whether it resolves over time. If some resolvers are still serving the old value well past your TTL, look for a different cause: a stale record left on one of your name servers, two name servers publishing different zone data, or a CDN or registrar-level override you had forgotten about.
Lower the TTL before you make the change, not after. Drop it to 300 seconds a day ahead, make the change once the old long TTL has expired everywhere, confirm the result here, then raise the TTL back up.
Lowering the TTL at the same time as the change does very little, because resolvers holding the old record also cached the old TTL along with it.
Each resolver is reported as one of:
It is governed by the TTL on the record that was cached, not by a fixed global schedule. Once a resolver's cached copy expires it fetches the new value. TTLs commonly range from five minutes to 24 hours, so a change can appear almost instantly in one place and take a day in another.
They are serving a cached copy that has not expired yet. A resolver that fetched the record shortly before your change is entitled to keep using it until the TTL runs out. If it persists well beyond your TTL, check whether all your name servers publish the same data.
Not from outside. You cannot clear another network's cache. What you can do is plan ahead: lower the TTL a day before a planned change so resolvers hold the old value for a much shorter time.
No, and we would rather be straight about it. We query ten major public resolvers from a single server in the United States. Because those resolvers use anycast, we reach their nearest nodes, so this shows whether the major resolvers agree — not how a change has spread country by country. That would require probes in each region.
Google Public DNS (both addresses), Cloudflare (both), Quad9, Cisco OpenDNS, Level3/Lumen, DNS.WATCH, Comodo Secure DNS and AdGuard DNS — ten endpoints covering the resolvers most people and networks actually use.