Brother-Eagle for Oracle provides several key measurements of a database's health, efficiency, and performance. The following table provides a master topic index to all of Brother-Eagle's Oracle performance tips and expert advice.
|Metric||Short Description (Click Metric Name for details)|
|# Active Users||
This parameter measures the number of active users. While there's nothing inherently wrong with a particular number of users, it is important to note heavy fluctuations in end user connections and activity.
|# Blocks Read/Sec||This parameter should be used to identify spikes in activity based on what is normal in your database. For instance, if 3.5 blocks/sec is normal read activity for your database, then you can set your thresholds accordingly to signal alarms above or below this number.|
|# Blocks Written/Sec||Like the number of Blocks read per second, this parameter identifies spikes in activity. Identify the normal write activity for your database in blocks/sec, then set your thresholds accordingly to signal alarms above or below this number.|
|# Broken Jobs||
This parameter indicates the number of Oracle jobs that are broken.
|Buffer Busy Waits||
Buffer busy waits occur when two processes are attempting to access the same buffer concurrently.
|Buffer Hit Ratio||
The buffer hit ratio has proven to give some insight into the performance of the database buffer cache. The idea is that if my query can access data directly from the buffer, then I avoided an expensive visit to disk. There's a lot that goes into determining the optimal buffer size but at the end of the day, not all of the data in the database can reside in memory.
|Dictionary Hit Ratio||
The dictionary hit ratio provides a general indication of how well the dictionary cache is sized for your database. Hit ratios should be used sparingly and should not be the main performance metric reviewed. Ratios should simply be used as performance indicators which may warrant further investigation into other areas.
|# Free Buffer Waits||
Free buffer waits occur when a server process was unable to find a free buffer and has asked DBWR to make a free buffer by writing out dirty buffers.
|Library Cache Hit Ratio||
The library cache hit ratio provides a general indication of how well the library cache is sized for your database.
|Locks||Locks commonly occur in an Oracle database; most of them are part of normal processing. The locks detected by this parameter are those that have remained in a lock state for at least a half second. Any locks that show up are an immediate cause for further analysis.|
|# Long Running SQL||#LongRunningSQL shows the number of statements in V$SQL that a long time to execute. A long running SQL statement is one that has been executing for more than 30 seconds. These statements should be scrutinized and are candidates for SQL tuning.|
|% Memory Sorts||This parameter indicates the percentage of disk sorts that are done in memory.|
|SQL Parse Ratio||This parameter represents the percentage of SQL statements in V$SQL that are being parsed multiple times. Ideally, you want the statement parsed once and executed many times|
|# Redo Log Space Requests||Redo log space requests occur when the active archive log is full and Oracle needs (waiting) disk space for the new redo entries.|
|%Table Scans (Short Tables)||This parameter depicts the percentage of full table scans that are performed on short tables. An ideal value would be 100%. While table scans are sometimes the desired optimizer execution path, they can present a tuning opportunity, especially when performed on long tables|