Welcome to Spotlight on IBM DB2 LUW (Linux, Unix, and Windows). The help topics in this book cover features available in the interface when Spotlight is connected to a version 9 database.
Spotlight is a powerful database monitoring and diagnostic tool. Its unique user interface provides you with an intuitive, visual representation of the activity on the DB2 databases and database partitions that you choose to monitor. Graphical flows and line and fill graphs illustrate the activity on and between database components. Icons display the value of key statistics. Using the Spotlight browser, you can switch your monitoring focus between the various databases and partitions. Additionally, you can view activity at the DB2 instance level for any of these systems.
The power of Spotlight lies in its ability to provide visual and audible warnings if the performance metrics exceed acceptable thresholds. The components and dataflows on the home page change color to show you the source of the problem.
A range of reports and graphs provide you with detailed information about a DB2 database, database partition, or the instance to which the database belongs. This information can be viewed on the screen or be printed.
You can set Spotlight options to warn you when a threshold is reached. You can define a number of thresholds so that warning messages are displayed well before the traffic levels into or out of databases become critical.
Spotlight uses a number of different techniques to warn you when your DB2 instance or a database is exceeding a threshold. For example, to issue a warning, you can configure Spotlight to change a color on the home page or drilldown, provide an audible signal, or perform an action, such as send an email message.
The following is a list of predefined alarms for the instance home page. To the right of each alarm name is the home page component and metric for which the alarm thresholds are defined. If necessary, you can use the Metric Editor to adjust the thresholds that correspond to these alarms to reflect your DB2 instance environment.
|
Alarm |
Home Page panel |
Component |
|---|---|---|
|
Global Sort/Join |
Post Threshold Sorts | |
|
Global Sort/Join |
Post Threshold Hash Joins | |
|
Global Sort/Join |
Piped Sort Rejects | |
|
Global Sort/Join |
Sort Overflows | |
|
Global Sort/Join |
Hash Loops | |
|
Global Sort/Join |
Hash Join Overflows | |
|
Global Sort/Join |
Hash Join Small Overflows | |
|
Global Application/Agent |
Agent Request Overflows | |
|
Global Application/Agent |
Agents Pending Locks | |
|
Global Application/Agent |
Agents Pending Token | |
|
HADR |
Missed Beats |
The Agents Pending Locks alarm becomes active when the number of agents waiting on a lock in comparison to the total number of registered agents for the DB2 instance reaches a threshold percentage. Lock waits occur for a number of reasons, including internal database contention.
This alarm has been predefined with a Low severity level. The underlying metric for this alarm can be adjusted to suit your environment by using the Metric Editor.
The application agents could be experiencing concurrency problems. Try to identify the applications that are holding locks for long periods of time and determine the cause. One possible solution is to use small units of work (for example frequent COMMITs) to reduce concurrency issues.
Access the Client Application Analysis drilldown to identify the applications holding locks and review statistics about those locks.
Notes:
The Agents Pending Token alarm becomes active when the number of agents waiting for a token (a permission to run) in comparison to the total number of registered agents for the DB2 instance reaches a threshold percentage.
This alarm has been predefined with a Low severity level. The underlying metric for this alarm can be adjusted to suit your environment by using the Metric Editor.
Determine whether the current value for the MAXAGENTS database configuration parameter is appropriate for your system. Some environments might use the MAXAGENTS parameter value to limit the number of processing tokens, thereby limiting the amount of execution concurrency in the database manager. Although this strategy might be beneficial to the system as a whole, it can impact applications, causing problems such a high percentage of agents awaiting tokens.
© ALL RIGHTS RESERVED. Termini di utilizzo Privacy Cookie Preference Center
