i try # read_timeout = “10s” and # write_timeout = “10s”, and stop questdb but telegraf still listen on port 8086 and my java app still think it is up!
Inputs and outputs run independently of each other. If an output goes down, telegraf will continue to pull metrics from inputs and buffer them. This way having one down output will not stop the collection of metrics.
The error should not come from the input, but the output.
As they run independently not really. You could build your own input plugin and use the exec input, which could check for the connection and stop sending metrics.