The Integration between OEM to SOI Connector via SNMP traps are sporadically getting lost and no alerts raised. We are missing critical table space issues alerts being caught. The OEM itself states that the SNMP traps are sent successfully.
The missing SNMP traps are discarded by the customized snmp_policy_oem.xml policy. We looked in to this policy and found a section to discarded traps "<!--Anything other than the specific Oracle trap types above will be discarded. --> <Field input='snmp_varbindvals' pattern='.*' output='eventtype' outval='Alert_Discarded' />" - In the ei-transform.log file search for Alert_Discarded and noticed it is the missing traps that are getting discarded. - In the SNMP policy any pattern matching under the following section will be published "<!--Add classification rules for specific trap types here-->" and the rest will be discarded. - To resolve this problem the pattern for the missing traps needs to be added in to the policy - Adding following section in the policy (under "<!--Add classification rules for specific trap types here-->") should resolve discarded traps problem for pattern "Pluggable Database", this is an example. This'll have to done for any pattern not matching. <Field input='snmp_varbindvals' pattern='.*, Pluggable Database,.*' output='eventtype' outval='Alert_CI_ PluggableDatabase ' /> - Once the changes are made in the policy do the following: Stop the catalyst container service Delete all files from \CA\Catalyst\CatalystConnector\container\nls-store and from \CA\Catalyst\CatalystConnector\container\registry-cache Start the catalyst container service.