Saving processor.dedup state has been added via pull (no release yet) which is great but led me to think what will happen if telegraf process is hard killed, crashes or the server crashes? This would be bad for me as I am looking for idempotence.
Is there a way to have timed state dump to ensure minimal data loss? I could do “kill -HUP” but this would effectively reload the entire telegraf process. Also see my other post. Is there some external signal that could be sent per processor to do something (like dump state, reload lookup tables etc).