Issuing snmp alarms in weblogic

Question

Hi,
I think there is a way for WebLogic (or EM FMC Control) to be configured to generate an SNMP alarm when there are app or managed server problems ?
Regards,
Harsha

Answer 1

To communicate with the WebLogic SNMP agent, you need to load the WebLogic Server management information base (MIB)
data into the MIB Browser. Using the MIB Browser application's File menu, select the Load MIB menu item, browse to the
${WL_HOME}/server/lib directory, and select the BEA-WEBLOGIC-MIB.asn1 file. After expanding the BEA-WEBLOGIC-MIB folder
on the left, you should see a list of WebLogic SNMP MIB tables.
This is interesting, but the main reason to use SNMP is to send unsolicited messages to the SNMP manager whenever something
happens. These unsolicited messages are called SNMP traps. WebLogic SNMP can generate traps to notify the SNMP manager
of certain types of events. WebLogic Server comes with a set of predefined traps for server startup, server shutdown, cold start
(admin server startup), and authentication failure. You can also set up three other types of traps: attribute change traps,
log message traps, and monitor traps.
The first step is to configure the Trap Viewer to listen for traps. Using the MIB Browser's View menu, select the Trap Viewer
menu item. Use the Trap Viewer's Start button to tell it to start listening for traps on its default port, port 162, with a Community
of public. Now, you need to configure the WebLogic Server side. For each SNMP agent, use the agent's Trap Destinations Configuration
tab to create a new trap destination. Set the Name to WebNMSTrapViewer, the Community to public, the Host to the IP address or
hostname of the machine where Trap Viewer is running (for example, localhost), and the Port to 162. If you are using SNMP v1.0
style traps, you do not need to specify the Security Name or Security Level attributes, which only apply to SNMP v3 style traps.
Do not forget to target the SNMP agent to the admin server.
Monitor traps are used to monitor an attribute value of an MBean; they come in three types: counter, string, and gauge.
A counter trap simply generates a trap when a particular attribute value meets or exceeds the threshold value. For example,
you might want to define a counter monitor trap to let you know when a server is using all of the connections in the
connection pool. To do this, you need to use the ActiveConnectionsHighCount attribute of the JDBCConnectionPoolRuntimeMBean
with the Monitored MBean Name of MyDataSourceName on ServerName.
A string monitor trap compares the attribute value against a string and can raise a trap when the string matches or when it differs.
A gauge monitor trap will alert you whenever the attribute value meets or exceeds the Threshold High value and when it reaches
or falls below the Threshold Low. If you have a JDBC connection pool where the Initial Capacity and Maximum Capacity attributes
are different, you might want to create a gauge monitor to monitor the maximum and minimum number of connections. By setting
the Threshold Low value to be one less than the Initial Capacity, your gauge monitor trap could monitor the
ActiveConnectionsCurrentCount attribute of the JDBCDataSourceRuntime MBean and alert you whenever the number of active
connections are less than the Initial Capacity (which might indicate database connectivity problems).