Discovering and managing database performance issues by using the Performance Dashboard
The IBM® BPM system
relies heavily on database queries. The more the system is used, the
more the database gets populated with artifacts, for example, process
instances and tasks, and the longer it takes for queries to return
results. You can identify potential performance issues
caused by large numbers of artifacts by viewing real-time and historical
artifact counts in the Performance Dashboard. To help you manage these
issues, the dashboard provides housekeeping tips for each artifact
type. If the artifact is associated with a snapshot, you can generate
the corresponding wsadmin commands to remove the snapshot from the
database. For offline analysis, you can export the data in the dashboard
to a spreadsheet.
Before you begin
- To view the dashboard, you must be member of the PerformanceAdmins team. For more information, see IBM Business Process Manager roles.
- To include database configuration and meta information in exported
data, you must have the following database access rights:
- DB2: select privilege on the DBCFG administrative view, SYSCAT.TABLES catalog view, and SYSCAT.INDEXES catalog view.
- Oracle: SELECT permission on the V$SYSTEM_PARAMETER view
- SQL Server
- The VIEW DATABASE STATE permission
- Select permission on the following system catalog views: sys.configurations, sys.stats, sys.tables, sys.objects, sys.schemas, sys.index_columns, sys.indexes, sys.sysindexes, sys.dm_db_partition_stats
- Permission to view metadata for all BPM tables and BPM indexes
About this task
- Process instances
- Current counts for each of the process instance states. The historical data shows the daily counts for started and completed processes. If process instances were deleted from the database, these instances are not reflected in the historical data.
- Tasks
- Current counts for each of the task states. The historical data shows the daily counts for received and closed tasks. If tasks were deleted from the database, these tasks are not reflected in the historical data.
- Groups
- Counts for groups by type and the 10 largest groups.
- Snapshots
- Counts for key performance indicators, such as named process application snapshots without active process instances, unnamed process application snapshots, and orphaned toolkit snapshots. Snapshot data includes system snapshots; do not delete these snapshots.
- Other
- Counts for various other artifacts, such as users, environment variables values, and Event
Manager tasks on hold.
It also includes data about the Process Portal index, such as the number of updated indexes and the index table rows that you can clean up. For more information, see Cleaning up index tables.