ERROR SolarWinds.AgentManagement.ServiceCore.

cjfranca

Guy I need help for undertend log file bellow.

"

2017-03-28 19:05:01,124 [31] ERROR SolarWinds.AgentManagement.ServiceCore.AgentManagementServiceApp - Agent 185 (node-01/172.17.35.19) has computer clock shifted by 05:04:56.1241149 from server time. Agent UTC time: 03/28/2017 14:00:05, server UTC time: 03/28/2017 19:05:01

2017-03-28 19:05:53,322 [69] ERROR SolarWinds.AgentManagement.ServiceCore.AgentManagementServiceApp - Agent 186 (node-02/172.17.35.30) has computer clock shifted by 05:10:37.3224208 from server time. Agent UTC time: 03/28/2017 13:55:16, server UTC time: 03/28/2017 19:05:53

2017-03-28 19:06:20,142 [82] ERROR SolarWinds.AgentManagement.ServiceCore.AgentManagementServiceApp - Agent 220 (node-02.buy4sc.local/172.17.35.225) has computer clock shifted by 05:04:56.1426523 from server time. Agent UTC time: 03/28/2017 14:01:24, server UTC time: 03/28/2017 19:06:20" 

So I have a problem with a larg memory and cpu used when the node poll is agent solarwinds, all server running windows, monitoring solarwinds Agent hight  use cpu and memory. But The server show hight process in varius different time.    The version agent is 1.6. 0  show in log,

jiri.tomek

Hello,

can you check the server CPU and memory directly using Task Manager to see which process/processes cause high memory and CPU consumption? Agent itself adds some overhead to the monitored node based on number of things you monitor there. If it's just CPU&memory is should be negligible but if you also monitor some applications the CPU and memory consumption may be higher. It's hard to tell what's the cause without knowing what you monitor via Agent.

If you monitor node via SNMP or WMI the "hard work" is on the Orion server, node just returns the raw data. That's fine if you have node on LAN and no network restrictions. But if you have nodes in different network with limited connectivity we need to limit the connections and data between node and Orion server and that's what Agent does. It executes the "hard work" locally on the node and only sends final results via well known channel. It simplifies monitoring of nodes behind firewalls but has higher demand on the node itself.

Another thing that I wee from log sample you sent is that your node clock is not synchronized with Orion server by a lot, over 5 hours difference. While that should not have direct impact on CPU and memory consumption it will cause data shift in all charts and other places in Orion. If agent has clock set to 1:00 and Orion server has 6:00, data that agent sends at 6:00 Orion time will appear in all charts as if they are from 1:00. I suggest fixing the clock on your servers to have meaningful results.