Stats mgmt is used to provide statistics about various elements within UCS. There are a few pieces to this. I'll do a step by step process
Threshold Policies
Each of the Server, LAN and SAN tabs have a Threshold Policy tab. These thresholds are used to generate events when specific thresholds are crossed. Normal value - this is the steady state and is a baseline used for the triggers.
Alarm Triggers (above) - the first check box is for the level of event to be generate. Then there are Up and Down values that are configured. When the value crosses the UP value, an event is generated. When the value crosses the DOWN value, the event disappears.
Alarm Triggers (below) - same idea as above except these are used when you want an event generated for below normal values (PS voltage as example).
Note
You cannot create or delete a statistics threshold policy for Ethernet server ports, uplink Ethernet ports, or uplink Fibre Channel ports. You can only configure the existing default policy. |
One thing worth noting is that there is some math involved determining what the UP/DOWN values should be. Here is an example for Ethernet:
The default collection interval is 1 minute and the reporting interval is 15 minutes. In order to make this as easy to read as possible, lets change those on the Adapter to 30 seconds and 2 minutes (the lowest values.
Lets says that we want a Major event if utilization of a 10G link exceeds 80% on the transmit side. We will use the 'ether tx stats total bytes' property.
Since this is in bytes we divide by 8 - 10,000,000,000/8=1,250,000,000 bytes
Multiple by 30 (collection interval) = 37,500,000,000
Now multiply that by .8 to get 80% = 30,000,000,000
Lets have the event clear if the utilization goes below 70% - 37,500,000,000 x .7 = 26,250,000,000
Here is a screen capture of the policy from the Admin Tab