eG Monitoring
 

Measures reported by PgDBStatsTest

For each database on the PostgreSQL server, this test reports the transaction load on the database and reveals how well the database processes the transaction requests to it and how well it utilizes its cache. Overload conditions and processing bottlenecks are thus revealed. The measures made by this test are as follows:

Measurement Description Measurement Unit Interpretation
DB_size Indicates the current size of this database. KB  
Cache_hit_ratio Indicates the percentage of requests to this database that were serviced by the cache, without having to read from disk. % Because reading from the cache is less expensive than reading from disk, you want the ratio to be high. The higher this value is, the better. Generally, you can increase the cache hit ratio by increasing the amount of memory available to the database server.

The detailed diagnosis of this measure provides you with the complete details of the database such as the number of server processes running on it, the number of transactions committed and rolled back, and the number of rows inserted, updated, and deleted.

Commit_ratio Indicates the rate at which live rows are fetched while this index is scanned. %  
Server_process Indicates the number of processes that are currently running on this database server. Number  
Inserted Indicates the rate at which the records are inserted into this database. Inserts/Sec  
Deleted Indicates the rate at which the records are deleted from this database. Deletes/Sec  
Updated Indicates the rate at which the records are updated into this database. Updates/Sec  
Commited Indicates the transaction throughput. Commits/Sec A decrease in this measure during the monitoring period may indicate that the applications are not doing frequent commits. This may lead to problems with logging and data concurrency.

The cause has to be probed in the application.

Roll_Backed Indicates the rate at which rollbacks occurred on this database. Rollbacks/Sec A high rollback rate is an indicator of bad performance, since work performed up to the rollback point is wasted. The cause of the rollbacks has to be probed in the application.
Blocks_read Indicates the rate at which the blocks are read from this database. Reads/Sec  
Blocks_hit Indicates the rate at which the blocks are fetched after a read is performed in this database. Hits/Sec  
Row_fetch Indicates the rate at which the rows that are read based on a user query are stored in the buffer. Fetches/Sec  
Row_return Indicates the rate at which the rows are fetched from the buffer and sent to the client application. Returns/Sec If the size of the rows that are fetched from the buffer is too large, then the rows are fragmented and transferred to the client which is time consuming. This may inturn affect the performance of the database to some extent.