In the Workflow editor, clicking Take action or Write message opens the
Action Settings window. Use this function to send a command or universal message to the system. For
example, you might want to log information, trigger an audible beep, or stop a job that is using too
many resources.
- Action Selection
-
Universal Message, when selected, opens this area:
- Category of the message you are defining. Type one of the following
categories: critical, warning, or informational.
- Severity that you want assigned to the message. Type a severity number
(1, 2, and so forth).
- Message that is displayed. Type the message text you want to be
displayed at the system specified (see Where should the action be run, below).
Universal messages are shown in the Universal Message Console
view.
-
System Command, when selected, opens a text box for entering the command to
issue at the system. The command language is that of the application and operating platform you are
running (such as a UNIX shell command to run at a UNIX system).
- Example: Windows has a command for sending a message to a system: net send computer_name
message, as in net send MYPC Disk is 85% full. Delete some files.
- Example: UNIX has the mail command for directing messages: mail node userID text.
- Attribute
- A window is opened for choosing an attribute to include in the
universal message or system command. In the Select Attribute window,
specify an attribute group and item, then click OK.
- The attribute you selected is displayed in the Message field
or the System Command field. The attribute
name is replaced by the actual value during execution of the action.
The attributes available for you to select are based on the situations
you have added to the policy preceding the action. If there are no
situations in the policy, the Attribute button
is disabled.
- In the example that follows, the command sends an email that reads,
"System: MyServer, Disk: D:, Space Available: 15%":
echo System: &NT_Logical_Disk.Server_Name;, Disk:
&NT_Logical_Disk.Disk_Name;, Space Available:
&NT_Logical_Disk.;%_Free%|mail userID.
- More Options
- If multiple data items are returned
- These options determine what action to take if more than one monitored
item of data is returned by the situation evaluation:
-
Only take action on first item issues the universal message or the system
command on only the first item of data that meets the situation criteria.
-
Take action on each item from activity issues the universal message or system
command on all items of data generated by the situation specified in the activity selected from the
list. The ID
for an activity is shown in the Grapher View when you move the mouse pointer over the activity
graphic.
- The results data returned from each of these individual situation-based
activities might originate from the same attribute group on an agent
or from attribute groups on different agents. For example, the workflow
might be triggered by an OMEGAMON XE for CICS on z/OS situation that
monitors for delayed transactions on an CICS monitoring agent, (activity
ID WaitOnSituation1) and might then evaluate the state of a
situation that monitors threads on an OMEGAMON XE for DB2 monitoring
agent, (activity ID EvaluateSituation2).
- Some situations can return a single row and others can return
multiple rows for each sample. You can control the number of universal
messages or system commands to issue when there are multiple situation
rows: a message or a command can be sent for the first row only or
for every row. As mentioned earlier, though, there can be multiple
sets of situation data originating from different activities. Select
this radio button to control which set of results data to use.
- Referring to the earlier example, send one command for every row
returned by the CICS situation by selecting WaitOnSituation1;
or send one command for every row returned by the DB2 situation by
selecting EvaluateSituation2.
-
Continue processing rows on failure, when selected, keeps track of both the
rows for which the command failed and those for which it succeeded. A policy can be constructed so
that one path is taken for the rows that succeeded and another for the rows that failed.
- This is a useful feature when, for example, you want to run a
script to filter results. In this case, a failed condition does not
mean that the activity itself encountered a problem but that a particular
row did not satisfy some criteria. When this option is cleared, a
system command that returns a non-zero result while processing a series
of situation result rows stops further processing of any remaining
rows and the activity is terminated with Error, even
though some of the commands might have succeeded.
Note: If the situation has a Display Item
set and the "Only take action on first item" option is selected, the
action is taken once for each unique display item in the rows of data.
For example, if there were 100 total rows that qualified and there
were 6 unique display items, the situation would fire 6 times. See
Formula for more information on display
items.
- Where should the action be executed
- Determines where to perform the action: on the system where the
agent resides, the monitoring server to which it is connected, or
the system selected from the list.
-
Execute the action at the agent that returned data for activity issues the
universal message or system command on the agent that returned data for the situation in the
specified activity (selected from the list). The ID for an activity is shown in the Grapher View when you move the
mouse pointer over the activity graphic. Referring to the Take action on each item from
activity example, if the action must be executed at the DB2 agent, you would specify
EvaluateSituation2 because this is the ID of the activity that evaluates the state of the DB2
situation on the DB2 agent.
-
Execute the action at the Server issues the universal message or system
command at the monitoring server that the agent connects to.
-
Execute the action at is used to select a specific managed system from the
list.