hard20 pts · first answer countskubernetesdnsperformance

Slow external DNS in Pods

Calls from Pods to api.partner.com are slow, and CoreDNS logs show many NXDOMAIN answers for that name. The Pod resolv.conf is below. What is happening?

search shop.svc.cluster.local svc.cluster.local cluster.local eu-west-1.compute.internal
nameserver 10.96.0.10
options ndots:5

A.With ndots:5 the name is tried with every search domain first, causing failed lookups; use a trailing dot or lower ndots.B.CoreDNS blocks external names unless they are listed in its ConfigMap, so each lookup retries until it times out.C.The nameserver IP is wrong; Pods should use a public resolver like 8.8.8.8 to resolve external names quickly.D.ndots:5 limits each Pod to 5 DNS queries per second, so extra lookups are queued and answered with NXDOMAIN.

← Throttled at 30 percent CPU Provider checksum mismatch in CI →