Performance report

[pavleec]

Hi,

Just wanted to share with others and thank you @toni-moreno for such a great tool :) So far I have over 515 devices out of which 320 is polled every 60s (145k fields) and no issues, gaps identified so far! Just being curious what are you're setups?

Regards,
PaVliK

[hyberdk]

We run 2855 devices, we collect between ~140 to ~15000 metrics per device. Not sure how I can check the total number of metrics. We also collect every 60 secs. This is done on a virtual RedHat enterprise. We run snmpcollector, influxdb and grafana on the same guest, all in its own docker container.

it has 8 CPUs, 16G mem and 1TB storage (need to add more storage soon)..

We do have small issues with spikes of PetaBytes in the graphs, but have not yet had time to look too close at it. it seems to be performing quite well..

Esben

[toni-moreno]

Hi (again) @hyberdk , I my last project we had 3 different instances each one for different teams, the biggest one, gathered metrics from 800 devices , about 450k metrics per minute.

You can check the metrics per device or per output db if you activate the selfmonitoring option as explained here(https://github.com/toni-moreno/snmpcollector/wiki/Troubleshooting#self-monitoring) and you can download and install two grafana dashboards to see internal performance.

[toni-moreno]

Hi @hyberdk did you test the our snmpcollector performance dashboards for grafana?

[hyberdk]

Hi @toni-moreno
I just enabled it, damn its slow to browse when there are lots of devices.. I can hear my CPU fan going nuts..

here is a screenshot:

it looks like I have ~400 devices that does not connect anymore.. Guess I need to get that device update script running every night ;-)