Scalability

[kotyara85]

👋 everyone,
Looking into the arch diagrams, it looks like basically every cluster will announce itself to the upstream DNS server.
What happens if you have a thousand clusters? 1000 NS records will get pushed to the upstream?

[abaguas]

Correct. This should not be a problem, the DNS infrastructure is robust and can handle it.
Each server will then query every other server to decide which records to publish. This operation will take some time, so a failover interval of 1s won't be feasible. But the same would apply with an external probing solution