I have a setup where we're running telegraf
and the pod is configured as a daemon set in rancher (so instance per node.) Now when doing nslookup
over the alias in any container, dns resolution returns all ips of all instances on all nodes. Is it possible to restrict this to just the node the workload is running on?
Reason to ask is that in some containers, there is a udp socket send_to, which for some reason seems to resolve all ips and then attempt to connect to all of them before writing the data.