TTL Tuner
Tune Your DNS for Speed, Reliability, and Control
Knowing what your TTLs are is the easy part. Knowing what they should be — and why — requires context. This tool detects your provider’s restrictions, explains the reasoning, and gives you exact copy-paste instructions to fix what’s off.
Understanding TTLs: What They Do and Why They Matter
What is a TTL?
TTL (Time to Live) is a value in seconds attached to every DNS record. It tells DNS resolvers how long to cache the record before asking the authoritative nameserver again. A TTL of 3600 means “cache this answer for 1 hour.” After that hour, the resolver discards it and fetches a fresh copy. Defined in RFC 1035 §3.2.1.
How TTLs Affect Page Load Speed
When a visitor’s DNS cache expires, their browser performs a fresh DNS lookup to find your server’s IP address. If the TTL is too short, that expiry happens more often — adding 50–200ms per fresh lookup. A TTL of 300s (5 min) means 288 lookups per day per resolver. A TTL of 3600s (1 hour) means just 24. That’s 92% fewer DNS queries and faster page loads for returning visitors.
The TTL Tradeoff
Longer TTLs = fewer queries, faster repeat visits, but slower propagation when you change the record. Shorter TTLs = faster propagation, more queries, slightly slower repeat visits. The right choice depends on how often you change each record type.
Common Mistakes We See in the Field
- TTL set to 60s on a production A record that hasn’t changed in years — generating thousands of unnecessary DNS queries per day.
- TTL set to 86400s (1 day) before a server migration — then wondering why the new IP takes 24 hours to propagate.
- Not realizing Cloudflare overrides TTLs on proxied records — setting TTL to 3600 but Cloudflare serving 300 regardless.
- Forgetting to raise TTL back after a migration — leaving it at 60s permanently, wasting resolver resources.
RFC References
RFC 1035 §3.2.1
Defines TTL as a 32-bit unsigned integer specifying the maximum time a resource record may be cached.
RFC 2181 §8
Clarifies TTL handling: resolvers must count down the TTL and discard records when it reaches zero.
RFC 8767
Serving Stale Data to Improve DNS Resiliency — defines behavior when TTLs expire during resolver failures.
