We are using DNS only (not proxied) load balancer with 10 minutes TTL for proximity based traffic steering.
Before enabling load balancer we had simple DNS records, not proxied, with the same 10 minutes TTL.
After enabling load balancer we noticed ~6x increase on DNS queries. It was unexpected because we calculated load balancer budget based on DNS queries.
Changing TTL on load balancer seems to have no effect on usage. Why?
Traffic before and after enabling load balancer with 10 minutes TTL: