Why DNS Propagation Takes So Long (And How to Know When It's Done)

You updated your DNS records. Your registrar's dashboard shows the new values. And yet, an hour later, your site is still pointing at the old server — or some visitors see the new one while others don't.

This isn't a bug. It's how DNS is designed to work. Here's what's actually happening and how to know when it's safe to move on.

DNS is a distributed system

There is no single DNS server that every computer in the world queries. Instead, DNS is served by thousands of resolvers — servers run by ISPs, cloud providers, offices, and individuals — that answer queries on behalf of their users.

These resolvers don't look up your domain fresh every time someone visits your site. They cache records for a period of time, then discard them and fetch fresh ones. That cache duration is set by you: it's the TTL (time to live) on each DNS record.

TTL is what controls the delay

When you set a TTL of 3600 on an A record, you're telling every resolver in the world: "Hold onto this answer for one hour before checking again."

That means:

• A resolver that just cached your record has up to an hour before it will see your change
• A resolver that cached your record 55 minutes ago will see the change in 5 minutes
• A resolver that hasn't seen your domain before will get the new record immediately

This is why propagation feels uneven — different resolvers are at different points in their cache cycle. Someone on a residential ISP might see your old record while someone on a corporate network already has the new one.

How to speed it up

Lower your TTL before making a change, not after. Once a resolver has cached a record with a TTL of 3600, lowering the TTL to 60 does nothing — the resolver will hold the old record for the full hour it cached it with.

The practical pattern:

1. Lower TTL to 300 (5 minutes) at least an hour before your planned change
2. Make your DNS change
3. Wait 5–10 minutes
4. After the change is stable, raise TTL back to 3600 or higher

The 48-hour myth

You've probably seen hosting companies say "DNS propagation can take up to 48 hours." This is a legacy figure from a time when DNS caches were longer and infrastructure was slower.

In practice:

• Most resolvers cache records for exactly the TTL you set
• With a TTL of 300, you'll typically see global propagation in under 10 minutes
• The worst case is a resolver that ignores your TTL and holds records longer — but this is increasingly rare

48 hours is possible with very high TTLs or misconfigured resolvers, but it's not a reason to wait anxiously when you've set a short TTL.

How to verify propagation

You can check what specific resolvers currently see for your domain:

• dig — query a specific nameserver: dig @8.8.8.8 example.com A
• nslookup — works on Windows too: nslookup example.com 1.1.1.1
• Online tools like dnschecker.org show results from multiple locations at once

The key is checking multiple resolvers, not just one. Your own machine's local resolver may be cached differently from what a user in another country sees.

Monitoring for unexpected changes

Propagation also cuts the other way: if something changes your DNS records without your knowledge — a misconfiguration, an expired record, or something more serious — you won't notice unless you're watching.

OneDollarDNS monitors your records by querying your authoritative nameservers directly and alerting you the moment anything changes, so you find out immediately rather than when a user reports a broken site.