Observability and Monitorability
Key components of modern-day enterprise software are observability (visibility into the activities occurring within the software) and monitorability (defining threshold limits for critical system metrics). IBM Planning Analytics provides both when using Planning Analytics Workspace (or PAW).
Monitoring Activity
You can view a detailed report of all current activities (known as “threads”) on any running Planning Analytics database. Using the information available in the report, you can cancel threads or disconnect users that are negatively affecting database performance.
When you open the Databases page in Planning Analytics Workspace, you can see a list of all databases running in your environment. If you click on “Threads” Planning Analytics attempts to authenticate you as an administrator and if your administrator credentials are authenticated, the database thread report is presented, providing details on all current activity.
In the thread report, each activity or “thread” will show it’s ID and Name as well as details such as: State, Type, Function, Wait, Elapsed ,W/R/Ix locks, Context, Info, Object Name and Object Type (you can read more about thread details here).
Thresholds and Alerts
You can configure threshold settings that define agent resource status on multiple metrics. When you set threshold limits, you define the parameters for the safe, concern, and critical states of these resources. You can also configure alerts to send notifications to designated individuals when any of the metrics reaches a critical state.
As of this date, the following metrics can be monitored by setting thresholds and configuring alerts:
- Memory Usage in GBs, as a percentage of installed system memory. A threshold can be set to indicate when memory usage has gone above expected amounts.
- CPU Usage as a percentage of total CPU capacity on the system. A threshold can be set to indicate when CPU usage has gone above expected amounts.
- Max Thread Wait Time measured in seconds indicates the average number of seconds any activity can run before potentially impacting performance.
- Thread in Run State measured in minutes, indicates the average number of minutes any activity can remain in a single state before potentially impacting performance.
- Database Unresponsive measured in seconds, should the Planning Analytics database become unresponsive, this is the maximum number of seconds allowed before taking corrective action.
- Database Shutdown this indicates that the Planning Analytics database has had an unplanned shutdown.
Metric Visibility in Workspace
In Planning Analytics Workspace, resource states are displayed on the Databases page. A check mark in a green circle indicates a resource in a healthy state. A yellow triangle with an exclamation mark indicates a resource in a concern state. A slash in a red circle indicates a resource in a critical state. To learn how to configure thresholds and alerts go here.
If you would like more information on the above, or assistance in configurating thresholds based upon your environment and your specific Planning Analytics Databases, you can always contact QueBIT.