Thank you for your suggestion! Our Observium installation is virtualized using KVM on CentOS. Can you share what changes you needed to make? Did you have to make your changes on the Linux host or guest? Or was it a VMware nerd knob?
-=JGR
-----Original Message----- From: Michael obslist@smarsz.com Sent: Thursday, January 2, 2020 1:18 PM To: Observium observium@observium.org; Justin Radke via observium observium@observium.org; observium@observium.org Cc: Justin Radke justin.radke@directlink.coop Subject: Re: [Observium] Gaps in Graphs
It sounds like you are getting a large amount of udp traffic at one time.
We run our observium server as a vmware guest. We had to do a fair bit of tuning to get udp and rx buffers to a point where we weren't losing responses (& in your case, syslog messages).
Might be worth looking into.
Michael
On 3 January 2020 7:29:29 AM AEDT, Justin Radke via observium observium@observium.org wrote:
Hello, periodically I have "gaps" in my graphs which appears to be when Observium was unable to reach a particular host. This problem occurs on multiple hosts at the same time on different network segments and I have been unable to find any commonality except for the Observium server itself. I noticed that the gaps in the graph also exist at the same times for Observium's checks of localhost. Checking localhost of course demonstrates this is not a lack of network connectivity outside of the server.
At these times I also found these entries but I'm not sure if they are related since they are merely syslog messages but thought I'd mention it: Jan 1 21:39:24 voldemort rsyslogd-2177: imuxsock begins to drop messages from pid 27161 due to rate-limiting Jan 1 21:45:50 voldemort rsyslogd-2177: imuxsock lost 22 messages from pid 27161 due to rate-limiting root@observium(~) :ps auxw | grep 27161 root 27161 0.0 0.0 200848 6520 ? S 2019 15:32 /usr/sbin/snmpd -LS0-6d -Lf /dev/null -p /var/run/snmpd.pid
What would cause this and how can I solve this matter?
-=JGR