How we decide
isitdns answers one question, "is it DNS?", and tries to answer it the way you would in a terminal: by asking DNS the same questions from more than one place and only trusting a problem when independent networks agree. This page is the method, in full. No login, no tracking, free forever.
// what we measure
We isolate the DNS layer from everything around it. For each thing we watch, we check it in order: can we reach it, does the transport work (Do53, DoT, DoH), does it resolve, does the answer validate (DNSSEC), is the delegation intact, and is the service actually answering. The verdict names which of those rungs failed, so "it's DNS" comes with "and here is the layer," not just a red light.
// where we measure from
The default fabric is three independent networks: a probe in New York, a probe in Phoenix, and the global edge. We show you how many are reporting right now, and we say so honestly when one goes quiet. We do not inflate that number, and we never claim to see the whole internet. We see what we watch, from where we watch it.
A count of three real, corroborating vantages with receipts is worth more than a number you cannot audit. The fabric grows over time; the honesty does not change.
// the rule
We only call it when independent networks agree. One vantage seeing trouble is a "maybe," not a verdict, because a single path can be wrong for reasons that have nothing to do with the thing it is asking about. A confirmed problem needs agreement across networks that fail independently, so a fault inside one provider's own network cannot masquerade as the target being down.
// when we won't answer
If our own measurement goes stale, the page says "Can't say" instead of defaulting to a calm green. A monitor that refuses to show a verdict it cannot stand behind is more trustworthy than one that never blinks. Every answer carries how fresh it is, so you can see for yourself.
// incidents and strikes
An incident is one continuous run of something we watch being degraded or down. A strike is one logged incident. We count incidents, not bad seconds, so a long outage is one strike, not thousands. We attribute a failure to the DNS operator that failed, not the website that fell over with it.
// how we talk about it
We report what we observed, never an accusation. "We observed a corroborated DNS incident involving a provider from N independent networks" is a statement about our own instrument, not an official status for anyone. We use plain names, no logos, and we are not affiliated with or endorsed by any provider we name. We publish honest data only: real measurements, or an honest "no data," never a placeholder.
// your privacy
Cookieless. No visitor tracking, no profiles, no PII. We do not want to know anything about you. The site is free and always will be.