Start of Content Area

Object documentation Performance Database Locate the document in its SAP Library structure

Definition

This sub-monitor in the SAP/Oracle Database Monitor gives you an overview of the load and performance of the Oracle instance.

Use

You choose Overall Activity  ® Performance Database.

You can use this sub-monitor to see if:

·        The database load has changed recently

·        The database instance has load peaks

·        One database instance is more heavily loaded than other instances – useful with Oracle Real Application Cluster (RAC)

Structure

The following structure applies to the tabs Overview, Intervals, Peak 10-12, Peak 14-16:

Column

Description

Weekday

Day of the week for the snapshot

Date

Date of the snapshot

Time

Time of the snapshot

User Calls

Number of user calls such as login, parse, fetch, or execute

Recursive Calls

Internal SQL statement or SQL statement in PL/SQL statement

User / Recursive Calls

Ratio of user to recursive Calls

User Commits

Number of commits and rollbacks

Parses

Total number of parse calls (hard and soft). A soft parse is a check on an object already in the shared pool, to verify that the permissions on the underlying object have not changed.

Reads / User Call

Amount of logical reads per user call. Should be less than 20.

Logical Reads

Sum of "db block gets" and "consistent gets"

Physical Reads

Number of physical reads

Physical Reads Direct

Number of reads directly from disk, bypassing the buffer cache

Physical Reads Direct (LOB)

Number of LOB reads directly from disk, bypassing the buffer cache

Buffer Quality

Percentage of how many db blocks are found in the db cache and haven’t to read from disk.

Physical Writes

Number of physical reads

Table Fetch by RowID

Number of rows accessed by RowID, including all rows accessed using indexes

Table Fetch Continued Row

Number of times when second row piece of chained rows is fetched. A high number indicates that rows are chained.

Table Scans Rows Gotten

Number of rows accessed by all full table scans. This is not the same as the number of rows returned because only qualifying rows are returned.

Table Scans Blocks Gotten

Number of blocks accessed by all full table scans

Redo Blocks Written

Total number of redo blocks written

Redo Write Time (ms)

Total redo write time since database start in milliseconds

Avg. Redo Write time (msec)

Average time the LGWR needs to write the redo log information from buffer to disk in milliseconds

Buffer Busy Waits

Number of times block access failed because another process held the block in incompatible mode. If this statistic is over 10% of logical reads then use V$WAITSTAT to check contention.

Buffer Busy Waits Time (sec)

Total buffer busy wait time since database start in seconds

Avg. Buffer Busy Waits Time (msec)

Average time a session has to wait for the event “buffer busy waits” in milliseconds

Full Table Scans

Sum of full table scans for long and short tables:

·        Short table scans are against tables with 4 or less database blocks.

·        Long table scans are against tables with 5 or more database blocks.

Note the following about the tabs:

·        All tabs include load and performance data for the days where snapshot information is available.

·        The Overview tab displays the data accumulated since database start on a daily basis.

·        The Intervals tab displays the data for each day.

Note

To see the load and performance data for every snapshot on a certain day, double-click the desired day in the Overview or Intervals tab.

·        The tabs Peak 10 – 12 and Peak 14 – 16 show the load and performance data at the peak times of 10:00 to 12:00 and 14:00 to 16:00. You cannot double-click here.

End of Content Area