立即与支持人员聊天
与支持团队交流

Spotlight on Oracle 10.10 - Getting Started Guide

Welcome to Spotlight Install Spotlight Start Spotlight Spotlight on Oracle Spotlight on Oracle Data Guard Spotlight on Oracle RAC Spotlight on Unix Spotlight on Windows Troubleshooting: Connection Problems

{Predict for:} / Execution Versus Time

To open the {Predict for:} / Execution Versus Time page

  1. Open the SQL Performance page. Select a SQL statement from the SQL Performance Predictions grid.

  2. Click {Predict for:} / Execution Versus Time.

Cost per Execution versus Time

Cost per Execution versus Time shows how the average cost of the SQL statement per execution will change between now and the specified end-date.

The Cost per Execution value is calculated on the Cost Metric (for example, buffer gets) used for the projection.

Chart data points

Data Points Description
The raw data (blue data points) The actual data used to generate the chart.
New data (in a contrasting color) New data (when available) is data collected AFTER the raw data was analyzed.
The shaded region The Prediction Interval that is based on the raw data.
A vertical trend change line

If displayed, it shows the early data follows a different trend from the more recent data. Trend Changes The early data is not used in making predictions.

Note: You can set a Trend Change line via the chart toolbar.

Chart toolbar

Toolbar Description

Add Trend Change

Mark trend changes on the Predictive Diagnostics chart.

In the Add Trend Change window:

  1. Use the arrow buttons to position the trend change line on the chart. (The trend change line appears on the chart as you use the controls.)
  2. Add some text in the Description box to identify the trend change.
  3. Select Also Create... to add the same trend change to related Predictive Diagnostics charts.

Delete Trend Change

Remove trend change lines from the chart.

In the Delete Trend Change window, select the Delete? option that corresponds to the line you want to remove.

Accept suggested trend change

If Predictive Diagnostics notices a possible trend change it will mark it on the chart and prompt you to accept or reject it.

Reject suggested trend change

 

Reanalyze

Generate a new Prediction Interval on the chart from the raw data.

You can mark a trend change on the chart wherever you like, but the trend change is valid only in circumstances where it represents a REAL change in Oracle's handling of SQL statement executions, system bottlenecks, or database resources.

  • You can expect (and should mark) a trend change when you change the environment where the SQL statements are executed — for example, when you install a new version of Oracle, or when you upgrade the platform that hosts the database.
  • Other trend changes may be clear from a visual inspection of the raw data displayed in the chart. If you believe you can see such a change, you should mark it. (If you add a trend change line where NO real trend change exists, you will NOT change the nature of Spotlight's prediction, but you WILL make it less accurate, as Spotlight is using fewer data points to generate the prediction.)

Note: Trend changes in SQL Performance Predictions charts apply only to the specified chart for the selected SQL statement. Predictions on the execution of SQL statements are specific to the SQL statements themselves. If one SQL statement will perform poorly on the Oracle database as data volumes and SQL execution rates increase, the same does not necessarily apply to other SQL statements. Equally, a SQL statement whose performance scales adequately with respect to one cost metric (CPU time, for example) may not scale adequately with respect to another (Disk Reads, for example).

Cost Summary

The Cost Summary for the SQL statement contains this information:

  • The trend prediction for the SQL statement (for example, Increasing at an increasing rate).
  • The current cost of the SQL statement as a percentage of the total current cost.
  • The predicted cost of the SQL statement as a percentage of the total predicted cost at the end-date.
  • The increase in the cost of the SQL statement at the end-date as a percentage of the current cost.
  • The increase in the number of cost events (for example, buffer gets) for the SQL statement between now and the end-date.

 

Related Topics

SQL Performance

相关文档

The document was helpful.

选择评级

I easily found the information I needed.

选择评级