In addition to the predefined alert types that are provided
by the web console, you can create your own alert types.
Before you begin
Ensure that you have the
Can Manage Custom Alerts privilege
on the repository database. In addition, you must provide a user
ID with the correct permissions on the database to run the script
for the user-defined alert type.
- To run a script on a database, ensure that the user ID that runs
the script has the required privileges to run the commands that are
included in the script on the database. For Executable/Shell script
job types, the user ID must also have permission to connect to the
database server by using SSH.
By default, the user ID that is included
with the schedule is the user ID that is stored with the database
connection. If that user ID does not have the correct privileges on
the database server, you can assign another user ID to run the script
when you schedule the script.
- When multiple databases are targets for a scheduled script, you
must have the Is Database Owner privilege granted
for each of the selected databases.
About this task
In your user-defined alert type definition, you must specify
a script that returns an exit value. The exit value determines the
severity of the alert to generate. You can also specify a schedule
for running the script against one or more target database servers.
Procedure
To create and manage user-defined alert types:
- From the Open menu, click .
- On the Alert Types tab, specify a script for the user-defined
alert type.
Specify a script that can run on the database
or the database server. The script type sets the connection method
to the database or database server. See sample
scripts for user-defined alert types.
The supported script
types are SQL-only scripts and Shell/Executable scripts.
- SQL-only script
- SQL Only scripts are run on the database. To run the job, the
job manager connects to the database and runs the SQL commands that
are included in the job script directly on the database.
- Executable/Shell script
- Executable/Shell scripts are run on the database server. To run
the script, the system logs in to the database server by using SSH
as the user ID that is defined in the database connection, and then
runs shell commands directly on the server.
Important: To
run Executable/Shell scripts on a database, ensure that the user ID
that is used to run the script has permission to log in to the database
server by using SSH.
If you want to use DB2 commands to generate
an alert, you can run the DB2 commands by using the Executable/Shell
script type for the user-defined alert type. Ensure that your DB2
environment is set up and initialized before you run the script.
The
script must return an exit value to indicate the severity of the alert
to be generated. The valid return values are:
- 0
- Normal condition (that is, no alert to be generated)
- -1
- Generate a warning alert
- -2
- Generate a critical alert
Important: The web console does not
verify that the scripts that you enter are valid. Click Test
Script to run the script on a database or database server,
or use other methods to verify that the script is correct and that
it produces the expected results before you schedule the user-defined
alert.
- Schedule the script to run on the database or the database
server.
A script runs according to an active schedule.
A schedule defines when a script runs, whether the script is repeating,
and whether the schedule is limited in number of runs or in time.
The schedule also defines one or more databases on which to run the
script. A schedule can be active, or inactive.
When you schedule
a script on a single database, you can define the user ID that runs
the script. If you schedule a script to run on more than one database,
the script runs on each database with the user ID that is stored in
the database connection for that database.
- Click Save All to save the user-defined
alert type.