Subject: [influxdb] Re: Collectd CPU metrics counting in jiffies, how do I get percentage?


пятница, 6 марта 2015 г., 16:07:04 UTC+3 пользователь [EMAIL PROTECTED] написал:
Hello!
Has anything changed for this issue?
I'm trying to setup collectd + influx + grafana, and looks like getting wrong values for types like counter/derive types.

using collectd from Debian Jessie (5.4.1-6), influx package - 0.9.5
server hostname is nagios.local and my configs looks like:
root@nagios:~# cat /etc/collectd/conf.d/network-listener-0.0.0.0.conf
<Plugin network>
  Forward true
  <Listen "0.0.0.0" "25826">

  </Listen>
</Plugin>
root@nagios:~# cat /etc/collectd/conf.d/network-server-nagios.local.conf
<Plugin network>
  <Server "nagios.local" "27826">

  </Server>
</Plugin>
for influx, collectd section (almost default, changed port):
[collectd]
  enabled = true
  bind-address = "0.0.0.0:27826"
  database = "collectd"
  retention-policy = ""
  batch-size = 5000
  batch-pending = 10
  batch-timeout = "10s"
  read-buffer = 0
  typesdb = "/usr/share/collectd/types.db"

what I had is graph flying high in Grafana which means that stored values are absolutes, not deltas.

--
Remember to include the InfluxDB version number with all issue reports
---
To post to this group, send email to [EMAIL PROTECTED].
Visit this group at http://groups.google.com/group/influxdb.