IBM Cognos Dynamic Cubes

Chapter 5. Administering dynamic cubes 123

5.2.3 Monitoring cube state through metrics

When managing dynamic cubes, a good practice is to monitor metrics displayed for each

cube in the Metrics window (Figure 5-15).

Figure 5-15 Dynamic cube metrics values

In addition to overall values, some metrics also have the value collected in the last hour, which

helps you keep track of both historic averages and recent changes to the system.

The following metrics are of particular interest in most cube management scenarios:

򐂰 Average successful request time

This metric indicates the average time for a report to execute successfully. It is useful

when monitoring the performance of the server and back-end database. Slow request

times might indicate performance issues that need to be investigated.

򐂰 Cube state

This metric indicates the current cube state: Disabled, Stopped, Starting, Running, and

Stopping.

򐂰 Data cache hit rate

A higher cache-hit rate indicates better utilization of data cache and better query

performance.

򐂰 Last metadata load time

This metric shows how long the previous load took for metadata to be loaded or refreshed.

Disproportionately long times may indicate modeling or configuration problems.

򐂰 Last response time

This metric shows how long the most recent time took to execute any request against this

cube. Useful for detecting performance degradation problems.