Maria Transactions Test

Rollbacks are costly operations on the database. This test monitors the percentage of rollbacks happening for user transactions with a database instance.

Target of the test : A Maria Database server

Agent deploying the test : An internal/remote agent

Outputs of the test : One set of results for the target Maria Database server being monitored.

Configurable parameters for the test
Parameter	Description
Test Period	How often should the test be executed.
Host	The IP address of the Maria Database server.
Port	The port on which the server is listening.
Database	Specify the name of the database that is to be monitored on the target Maria Database server.
User and Password	The eG agent has to be configured with the credentials of a user who has server-wide PROCESS and SELECT privileges on the monitored Maria Database server. To know how to create such a user, refer to Configuring the eG Agent with Access Privileges section.
Confirm Password	Confirm the password by retyping it here.

Measurements made by the test
Measurement	Description	Measurement Unit	Interpretation
User commits	Indicates the number of user commits that have happened during the last measurement period.	Number
User rollbacks	Indicates the number of user rollbacks that have happened during the last measurement period.	Number	Ideally, there should be few user rollbacks happening. Typically, whenever a delete, insert or update operation is performed on the database, Undo tablespace is consumed, I/O overheads increase, and considerable server time is spent in performing that operation. When such operations are rolledback, these resources are wasted! To conserve resources, its best to keep rollbacks at a minimum.
Rollback percentage	Indicates the number of user rollbacks as a percentage of the total user transactions (user commits + user rollbacks) with the database	Percent	The closer the percentage of rollbacks is to zero, the lower the overhead on the database due to rollbacks. The acceptable value of rollbacks will vary from one instance to another and will have to be configured based on the patterns of requests being handled by the database instance.