DNS changes don't take effect everywhere instantly — understanding propagation and TTL (Time To Live) explains why, and helps you plan DNS changes to minimize disruption.
What TTL Controls
Time To Live specifies how long (in seconds) a DNS resolver is allowed to cache a record before checking for updates again — a higher TTL means better caching performance but slower propagation of changes; a lower TTL means faster propagation but more frequent lookups.
Why "Propagation" Isn't Instant
DNS relies on a distributed caching system — ISP resolvers, individual devices, and various intermediate caches all hold onto DNS answers for the specified TTL duration, meaning different users may see the old vs new record simultaneously until every cache's TTL expires.
Typical TTL Values
| TTL | Use Case |
|---|---|
| 300 seconds (5 min) | Planned migration window, wanting fast propagation |
| 3600 seconds (1 hour) | Reasonable general-purpose default |
| 86400 seconds (24 hours) | Stable, rarely-changing records, maximizing caching efficiency |
Planning a DNS Change: Lower TTL in Advance
Before a planned migration or DNS change, lower the TTL on the relevant record(s) a day or more ahead of time — this ensures caches expire and pick up the new value quickly once you actually make the change, rather than being stuck with a long-cached old value.
After the Change: Raise TTL Back
Once the change is confirmed stable, raise the TTL back to a more efficient longer value — reduces unnecessary DNS query load and slightly improves resolution speed for end users.
Checking Current Propagation Status
dig yourdomain.comRun from multiple locations/networks (or use an online propagation checker tool that queries from various global locations) to see how widely a change has actually propagated.
Why You Might See Different Results in Different Places
Different DNS resolvers (your ISP, a public resolver, your local device's cache) may have cached the record at different times, resulting in genuinely different observed states of propagation simultaneously — this is expected DNS behavior, not a sign of misconfiguration.
Common TTL Mistakes
- Making a DNS change without lowering TTL first, then being surprised by slow, inconsistent propagation
- Leaving TTL very low permanently, causing unnecessary DNS query overhead without an active reason
- Assuming propagation is complete after a short wait, when caches with long TTLs may still hold the old value for hours
NS (Nameserver) Record Propagation: A Special Case
Changing your domain's nameservers themselves (not just individual records) is subject to a separate propagation consideration at the registrar/root level, sometimes taking longer than typical record-level TTL — see How to Migrate DNS to a New Provider Without Downtime for handling this specific scenario carefully.
Common Errors
Change appears live for me but not for others — expected during the propagation window; verify with a multi-location DNS checking tool rather than relying solely on your own local resolution.
Continue Reading
- How to Migrate DNS to a New Provider Without Downtime
- DNS Fundamentals: A, AAAA, CNAME, MX, TXT & NS Records Explained
- How to Use dig and nslookup for DNS Troubleshooting
Browse more articles in Networking & DNS.
