ESC snmp get threshold wrong target OID

[neetimit]

Hi,

I am using ESC to deploy a CSR VNF on openstack and creating metric on ESC for CPU load threshold check to scale up and scale down. I have attached my ESC deployment xml. 

ESC shows in mona.log file that it is hitting OID: UCD-SNMP-MIB::laLoad (.1.3.6.1.4.1.2021.10.1.3.1) to get the CPU avg load for 1min from CSR.

this is from the logs: Threshold Value from Agent :udp:98.6.63.71//161forrule:rule-VM_OVERLOADED-austx-lab-dev-esc-01-vnf-info-csrscal8__929824d8-f97a-4979-bba1-d2fb8ac741c0__0 for .1.3.6.1.4.1.2021.10.1.3.1= 0.0

I checked on Cisco MIB locator tool that this OID which ESC is searching on CSR vnf is not supported by CSR1kv XE universal and also if I do snmpwalk from ESC server it says No such object exist, that's why it is returning 0.0 value always even though I see CPU load is higher.

[root@austx-lab-dev-esc-01 ~]# snmpwalk -v2c -c public 98.6.63.71:161 1.3.6.1.4.1.2021.10.1.3.1

UCD-SNMP-MIB::laLoad.1 = No Such Object available on this agent at this OID

The correct OID which ESC should search on CSR VNF is in CISCO-PROCESS-MIB (cpmCPUTotal1minRev): 

[root@austx-lab-dev-esc-01 ~]# snmpwalk -v2c -c public 98.6.63.71:161 1.3.6.1.4.1.9.9.109.1.1.1.1.7

SNMPv2-SMI::enterprises.9.9.109.1.1.1.1.7.7 = Gauge32: 0

My question: How do I tell ESC to pull CISCO-PROCESS-MIB (cpmCPUTotal1minRev) and not the UCD-SNMP-MIB::laLoad. 

Is there any snmp conf on ESC server where I can add this MIB? How do I load vendor specific MIBs?

Any help appreciated!!

Thanks,

Neetika

[ydorego · Accepted solution]

Yes as per the reported error, when applied to CSR the OID is incorrect because the OID used is the one associated to a linux MIB.

In oder for you to target a specific OID, you simply have to defined a new metric that will specifically retrieve that OID.

The following URL describe the steps to add a new metric/action. That new metric will make use of the new OID that you want to target. For example you can introduce a CSR_CPU metric.

 https://confluence-eng-sjc1.cisco.com/conf/display/ESCWIKI/Dynamic+Mapping+API

The link above has an example describing metrics for CSR OUTPUT_TOTAL_BIT_RATE.

Your new metric would be something like the following:

<metric>
        <name>CSR_CPU</name>
        <userLabel>CSR CPU</userLabel>
        <type>MONITOR_THRESHOLD</type>
        <metaData>
            <properties>
                <property>
                    <name>target_oid</name>
                    <value>.1.3.6.1.4.1.9.9.109.1.1.1.1.7</value>
                </property>
                <property>
                    <name>agent_address</name>
                    <value></value>
                </property>
                <property>
                    <name>agent_port</name>
                    <value>161</value>
                </property>
                <property>
                    <name>agent_protocol</name>
                    <value>udp</value>
                </property>
                <property>
                    <name>community</name>
                    <value>public</value>
                </property>
            </properties>
            <type>snmp_get_threshold</type>
        </metaData>
    </metric>

That new metric would have to be used instead of the default CPU one provided by ESC.

Thanks,

[neetimit]

Thanks very much! The link is pretty awesome to understand the KPIs.