MikroTik

Could someone help me out with setting it up so anytime the CPU % on my cisco switches goes over 50% that I get an alert?

Most of this is covered in the probe thread. Create a function to read the cpu and a probe to graph and notify... Function
Cisco_CPU_available
array_size(oid_column("1.3.6.1.4.1.9.2.1.57", 10, 29))

Probe
Cisco_CPU_available()
if(or(Cisco_CPU_available()<>"", Cisco_CPU_available()>95), "","CPU Problem")
oid("1.3.6.1.4.1.9.2.1.57.0", 10, 29)

These have been updated to cache the value for 10 seconds and negative cache the value for 29 seconds (Negative cache default is 300 seconds, Negative cache gets set if SNMP fails to read the value). Which means if your probe fails it will stay failed for 5 minutes. This works correctly with 30 second polling intervals.

The OID I use is for the 1 minute average. Since I am polling every 30 seconds might as well get the 1 minute instead of 5 minute or 5 seconds..

Hi, lebowski.

does it alerts you?
Cisco_CPU_available
array_size(oid_column("1.3.6.1.4.1.9.2.1.57", 10, 29))

if(or(Cisco_CPU_available()<>"", Cisco_CPU_available()>95), "","CPU Problem")

Cisco_CPU_available always will return 1.

The correct solution should be

<?xml version="1.0" ?>
<dude version="3.5">
<Function>
<sys-type>57</sys-type>
<sys-id>52242</sys-id>
<sys-name>Cisco_CPU</sys-name>
<code>if(array_size(oid_column("1.3.6.1.4.1.9.2.1.58")) > 0, oid("1.3.6.1.4.1.9.2.1.58.0"),"False")</code>
<descr>**Cisco cpu 5 minutes average</descr>
</Function>
</dude>

<?xml version="1.0" ?>
<dude version="3.5">
<Probe>
<sys-type>13</sys-type>
<sys-id>41288</sys-id>
<sys-name>Cisco_Cpu_5</sys-name>
<typeID>8</typeID>
<functionAvailable>Cisco_CPU() <> "False"</functionAvailable>
<functionError>if(Cisco_CPU() <> "False",if(Cisco_CPU() < 80, "", concatenate("Warning: hi CPU usage. Current CPU_5_Min_Average_Usage=", Cisco_CPU(), "%")),"Not a cisco device or SNMP access timeout")</functionError>
<functionValue>oid("1.3.6.1.4.1.9.2.1.58.0")</functionValue>
<functionUnit>%</functionUnit>
</Probe>
</dude>

Oh, that is great all this time, Yes you are correct sir, it doesn't report high CPU. Oh well THANKS for the fix.

On another note pasting XLM into the dude can break your maps.
You should only paste the lines of your probe/function and not the XML.

Thanks gsandul, I'll fix up my probe and repost it in the Probe thread...

I am trying to get your probe to work but I get a Parse Failed,
Lebowski

Alright that is great, Fixed my probe up with your code everything working.

Gsandul you should post more often
Thanks, Every time I have seen one of your solutions they have always been great.

Lebowski

Well we are back to square one, The probe fails if the cpu average is 0%. Tried a few things but no luck. I will report if I find a way to fix it.

I decided to cheat and add one, this will never report as 0 as well but it will report False and not fail when the cpu average is 0...
The graph is the actual value so this should solve all the issues.
I am also modifying cache and negative cache as usual.

if(array_size(oid_column("1.3.6.1.4.1.9.2.1.57", 10 ,29)), oid("1.3.6.1.4.1.9.2.1.57.0", 10, 29)+1 ,"False") And the probe...
if(Cisco_CPU_a()<>"False",if(Cisco_CPU_a() < 50, "", concatenate("Warning: high CPU =", Cisco_CPU_a(), "%")), "Failed read") If you want to use the function in a label you can just subtract 1 from it.
CPU: [Cisco_CPU_a()-1]

FINALLY