eG Monitoring
 

Measures reported by OraTranWorkloadTest

Knowing the count of transactions executing on the Oracle database server per second can indicate the transaction load on the server. However, the true impact of this load can be assessed and understood only if administrators are enabled to determine the number and type of database operations each transaction triggers. This is where the OraTranWorkloadTest test helps!

This test reports how many key database operations - eg., data modifications, block changes, reads/writes, parses, rollbacks, etc. - are performed on the server per transaction. This way, the test reveals the real workload of the server. In addition, the test also enables administrators to compare current CPU usage with the real workload, so that they can figure out whether/not the server needs to be resized to handle its load.

The measures made by this test are as follows:

Measurement Description Measurement Unit Interpretation
Redo_size Indicates the amount of data written to the redo logs per transaction since the last measurement period. MB/Trans If the value of this measure keeps growing, it could indicate that transactions are making numerous and frequent changes to the data in the databases.
logical_reads Indicates the number of logical reads performed by the server per transaction. Reads/Trans These measures are good indicators of the level of activity that every transaction generated on the database server.
Block_chgs_count Indicates the number of database blocks that were changed per transaction. Blocks/Trans
Phy_reads Indicates the number of physical reads performed per transaction. Reads/Trans
Phy_writes Indicates the number of physical writes performed per transaction. Writes/Trans
User_calls Indicates the number of user calls made per transaction. Calls/Trans  
Parses_count Indicates the number of parses executed by the server per transaction. Parses/Trans Parsing is one stage in the processing of a SQL statement. When an application issues a SQL statement, the application makes a parse call to Oracle Database. During the parse call, Oracle Database:

  • Checks the statement for syntactic and semantic validity.

  • Determines whether the process issuing the statement has privileges to run it.

  • Allocates a private SQL area for the statement.

If the value of this measure keeps increasing consistently, it could indicate on an average, transactions are executing many SQL statements on the server, thus generating more parses.

Hard_parses_count Indicates the number of hard parses executed per transaction. Parses/Trans As opposed to a soft parse, a hard parse loads the SQL source code into RAM for parsing. A high value for this measure therefore indicates that the server is performing many hard parses.
WA_mem_processed Indicates the amount of work area memory used up by the server per transaction. MB/Trans OOracle Database reads and writes information in the PGA on behalf of the server process. An example of such information is the run-time area of a cursor. Each time a cursor is executed, a new run-time area is created for that cursor in the PGA memory region of the server process executing that cursor. For complex queries (such as decision support queries), a big portion of the run-time area is dedicated to work areas allocated by memory intensive operators, including:

  • Sort-based operators, such as ORDER BY, GROUP BY, ROLLUP, and window functions

  • Hash-join

  • Bitmap merge

  • Bitmap create

  • Write buffers used by bulk load operations

For example, a sort operator uses a work area (sometimes called the sort area) to perform the in-memory sort of a set of rows. Similarly, a hash-join operator uses a work area (also called the hash area) to build a hash table from its left input. If the amount of data to be processed by these two operators does not fit into a work area, then the input data is divided into smaller pieces. This allows some data pieces to be processed in memory while the rest are spilled to temporary disk storage to be processed later.

A consistent increase in the value of this measure is indicative of excessive usage of the work area by transactions. This could indicate that the transaction workload is characterized by complex queries that use memory intensive operators such as sort, hash-join, etc. You may want to fine-tune the work area size in order to enable it to handle the memory-intensive load better.

Logons Indicates the number of users logging in per transaction. Logons/Trans A steady rise in this value is indicative of a steady increase in user activity on the server.
Executes Indicates the number of executes performed per transaction. Executions/Trans  
Trans_rollbacks Indicates the number of rollbacks performed per transaction. Rollbacks/Trans Ideally, the value of this measure should be low. This is because, rollbacks are expensive operations and should be avoided at all costs. A consistent increase in the value of this measure is hence a cause for concern.