Hi
I use telegraf to gather general metrics (e.g. CPU, disk usage, PostgreSQL stats, etc.) but I'm not using it to monitor my Tor instances.
Le 08/12/2018 à 04:01, petrarca@xxxxxxxxxxxxx a écrit :
> Questions I could not answer yet myself include e.g.
>
> - What will be the performance impact of running Telegraf on a Tor-Realy (CPU, disk, I/O,...)?
>
I don't think telegraf will have a big impact but that will depend on how many metrics you'll be gathering and at what frequency.
> - What happens if the InfluxDB (running on another system) is not reachable (maintenance, outages) - is the data lost or buffered somehow?
>
telegraf has a buffer mechanism Have a look at the configuration reference:
https://github.com/influxdata/telegraf/blob/master/docs/CONFIGURATION.md#agent-configuration
> - I would like to also keep see old/historical data but not as granular as more recent data - is there any way to boild down old data, e.g. to only keep hourly data for data older that a month etc.?
>
You could use InfluxDB's retention policies and continuous queries to do so.
--
Johan Fleury
PGP Key ID : 0x5D404386805E56E6
_______________________________________________
tor-relays mailing list
tor-relays@xxxxxxxxxxxxxxxxxxxx
https://lists.torproject.org/cgi-bin/mailman/listinfo/tor-relays