When we start CHSinker we get in console:
"Run http server http://10.42.2.253:2221"
Problem: after moving to Single Task per Process (in 1.8) we had to add image: vadv/prometheus-exporter-merger in order to consolidate all http metrics (for 20+ tasks) into one metric prometheus output. Consolidator (prometheus-exporter-merger) is running in same pod as all chsinker containers. We normally configure consolidator running in same pod as connecting to localhost.
example: value: 'http://localhost:2221/metrics,type:chs_chsinkertask1'
However in case of ChSinker we see that ChSinker binds to Specific IP:PORT not to anyIP:port
We are running ClickHouse_Sinker on kubernetes.
When we start CHSinker we get in console:
"Run http server http://10.42.2.253:2221"
Problem: after moving to Single Task per Process (in 1.8) we had to add image: vadv/prometheus-exporter-merger in order to consolidate all http metrics (for 20+ tasks) into one metric prometheus output. Consolidator (prometheus-exporter-merger) is running in same pod as all chsinker containers. We normally configure consolidator running in same pod as connecting to localhost. example: value: 'http://localhost:2221/metrics,type:chs_chsinkertask1'
However in case of ChSinker we see that ChSinker binds to Specific IP:PORT not to anyIP:port
Is there any special reason to have IP:port bindings (except that it is nice to see IP) and not general port only
(http.ListenAndServe(":2121", nil))