Page History
...
The Statistics tab is primarily intended for instance tuning. You can also use the tab to periodically monitor the health of your instance. It lets you provide answers to the following types of questions: Is Oracle using resources efficiently? Is the SGA correctly sized? Do our systems have enough memory to enable us to keep adding dedicated server processes? Are we performing too many logical I/Os?.
With each release of the database, Oracle introduces new statistics. For example, operating system-related statistics were introduced in Oracle 10g. Precise for Oracle displays this information when available.
...
If you want detailed information on a wait event, in the Association area, click the row of the event that you want to view detailed information for. The Tab heading indicates the newly selected entity, and the Main area displays over time graphs for the wait event you drilled down to. There is no Association Area data for a wait event. See “How How most tabs are structured” on page 22structured.
| Anchor | ||||
|---|---|---|---|---|
|
...
| Column | Description |
|---|---|
| Wait Event | Name of the Oracle wait event. |
| In Oracle Sub-State | Equivalent Precise for Oracle wait sub-state. See “Wait States” on page 35.* Wait States. |
| Idle | Indicates whether Precise for Oracle categorizes the wait event as an idle event. Precise for Oracle considers certain events, such as sql*net message from client, rdbms ipc message or pmon timer, as idle because Oracle is effectively sleeping and not consuming any resources. The categorization affects the top 10 wait events displayed in the Wait Event view of the Tier or an instance. |
| Waits | Total number of occurrences of the wait event. |
| Timeouts | Total number of occurrences of the wait event that exceeded the timeout. Some events have a timeout associated with them. |
| Times Waited (Sum.) | Total time waited, displayed as a duration and bar graph. |
...
You can use the More... option in the Association controls to populate the association area with a subset of statistics based on class. For example, you can click Statistics>Cache or Statistics>Debug Statistics > Cache or Statistics > Debug. Some statistics are classified more than once. For example, Buffer is Pinned Count is classified as both Cache and SQL.
...
| Info |
|---|
The pie chart will not display what are considered to be idle events. |
See “Examining latches” on page 176 Examining latches.
To examine top 10 wait events
...
One of the recommended ways to tune statements is to reduce the number of logical I/Os, because logical I/Os tend to dominate response time. It is important to understand how the number of logical I/Os varies over time. The instance statistics in the Statistics tab re-enforce the resource consumption figures observed in the Activity tab.
| Anchor | ||||
|---|---|---|---|---|
|
We can tune statements by examining the current number of consistent gets for a period of time.
...
- First examine logical I/Os. See “Examining consistent gets” on page 176 Examining consistent gets.
- Now examine physical I/Os. For our example, the Statistics tab shows sampled data from the same time period as that shown for consistent gets over time. We can see that the average read and write time follows the number of physical I/O operations reasonably close, particularly the first peak on September 22nd. This indicates that there is some contention between I/O operations, perhaps because they are all reading from different points on the same disk, causing a rise in seek time.
...
For our example, we will examine how a number of hard parses affects the parse time. If we look at the Hard Parse Ratio graph in the Main area, we can observe that the hard parse ratio increased to around 50% at 6 PM, meaning that some new statements were executed or required reloading. Consequently, the total parse time increased to around 15 minutes. Then the number of statements that required parsing remained fairly constant (shown on the Soft Parse graph), but the hard parses fell away and so did the parse times. This example explains why it is a good idea to use bind variables or cursor sharing.