I’m currently testing version 2.0.0 (fb95362) running on an AWS EC2 instance and generating the configuration file for Telegraf via the GUI. I noticed that it uses the IP address of the GUI’s URL whenever it generates the configuration file (e.g. http://54.xxx.xxx.xxx:8086).
For example, if I access it using the public IP address of the EC2 instance, it will use that public IP address when it generates the Telegraf URL configuration.
urls = ["http://54.xxx.xxx.xxx:8086"]
I like the idea of just pointing the Telegraf agent to a remote configuration file instead of generating the configuration file on the remote machines but is it possible to tell InfluxDB that I want to use a base IP address for the URLs instead of it using the IP from the URL used to access the GUI?
I think it would be worth opening a feature request on the influxdb repository for this, I can imagine many reasons this would need to be different such as when the write path is through a proxy.
Yea. 100% needs to be configurable. Why would I use a local IP when DNS exists? That way if a system fails or something needs to move, I don’t have to update the configuration on hundreds or thousands of telegraf (or other) collectors.
Also for HA, I may have a proxy in use. Something to provide TLS termination because everything should behind TLS - even locally.