We have a system that depends on Datadog’s Custom Emitters feature that reads in raw metric data, then allows for custom actions on them.
That feature was removed from the newer DataDog agent version, but we would still like to upgrade and be able to maintain a critical code path that was introduced via that emitter feature.
So, multi-part question,
- Can telegraf act as a proxy to collect our existing statsd inputs? (say, replace the dogstatsd listener on 8125 with telegraf, and start dogstatsd on another port). I see no UDP/TCP output plugin, so I’m guessing not. Edit - Can
outputs.socket_writerbe used here? If so, what output format should be used to passthrough statsd metrics as-is? - I see there is a datadog Telegraf output, but how is this maintained? I think I would trust the datadog-agent package installation rather than any custom client code is used in Telegraf
- I assume we can have something other than the exec output plugin to operate/transform/filter the statsd input before it would arrive at this “forwarded” statsd output?
