Telegraf detailed feature documentation


i am looking for detailed documentation of Telegraf.
I am trying to see what features exist. in particular i am looking for

  • retries when output fails
  • batching of data send
  • etc

is there some detailed documentation about telegraf somewhere that describes the way it is actually working?


@mantzas This glossary of terms should help you out, but let me try to explain as well.

  • Telegraf will scrape the telegraf.Accumulator instance in each plugin and pool the resulting data in the metric_buffer every interval
  • Telegraf will attempt to write batches of metric_batch_size to each configured output every flush_interval
  • Telegraf will cache points locally in the event that it cannot send to a configured output. That cache will not be larger than metric_buffer_limit. If the cache gets larger than that then points will begin to be dropped off the end of the metric_buffer

Hope this helps! Is that what you were looking for?

yes, this helps.

i just wondered how telegraf is working when a output is down.


1 Like