I have a 2-nodes cluster with this setup for the networking configuration :
http.host: [_local_,_ens192_] http.port: 9200 transport.host: [_ens161_] transport.port: 9300
And here is my network setup :
# ip -br a lo UNKNOWN 127.0.0.1/8 ens161 UP 172.19.4.230/24 ens192 UP 172.19.3.230/24
When I query the CAT API for nodes I get this :
# curl "http://172.19.3.230:9200/_cat/nodes" 172.19.4.231 38 96 2 0.00 0.01 0.00 cdfhilmrstw - host01 172.19.4.230 19 97 3 0.01 0.08 0.08 cdfhilmrstw * host00
The reported IP addresses are for the "transport" network and not the "http" network.
It seems weird since the API is mostly used to understand the state of the cluster and it prevents sniffing for clients or monitoring.
Is it normal?
Is there a way to change this?