Time to Live (TTL)
Why TTL Matters
TTL directly controls the tradeoff between DNS performance and agility. A high TTL (like 86400 for 24 hours) reduces DNS query volume and slightly improves page load speed because the resolver serves cached results. But it means DNS changes take up to 24 hours to propagate fully.
For sites that rarely change their DNS records, a TTL of 3600 (1 hour) to 86400 (24 hours) is reasonable. For records that need to change quickly - like those used for DNS-based failover or load balancing - a TTL of 60 to 300 seconds allows rapid updates. Services like Cloudflare often use automatic TTL management.
A common operational mistake is setting a long TTL and then needing to make an urgent DNS change (like pointing to a new server during an outage). With a 24-hour TTL, the old record will be served from caches for up to 24 hours. Best practice is to lower TTLs before planned changes and have a standard playbook for emergency DNS updates.
How to Check
A DNS health checker shows the TTL for each of your DNS records. Review whether the values are appropriate for each record type. Critical records that may need to change quickly (A records, MX records) should not have excessively long TTLs.
Quick Reference
| Scenario | Recommended TTL |
|---|---|
| Stable production records | 3600 (1 hour) to 86400 (24 hours) |
| Before planned DNS changes | 300 (5 minutes) |
| During migration | 60-300 seconds |
| CDN/load balancer targets | 300-600 seconds |
| MX records | 3600-14400 seconds |