FineBI Performance Optimization

When you use FineBI, the following issues may affect the performance:

• The loading time of accessing dashboards is too long.

• Too much server memory is occupied due to frequent access to dashboards of large amount of data, resulting in memory overflow.

• Server crashes caused by server overload due to concurrent access of too many users.

• Frequent request timeout.

• Long update time, update errors, or stuck updates.

If the above issues occur, FineBI performance needs to be improved for better usage experience.

When there are performance issues, you should first analyze the causes and then the solutions.

Server Performance

As a software developed by Java, FineBI that has been integrated to a server will inherit the resources of the server. The virtual memory of the server, settings of the connection pool, and so on often lead to many performance issues.

Data Preparation

• In self-service dataset creation, if there is Cartesian product when you join tables, update failure or long update time may occur because of data inflation.

• SQL dataset preview speed affects the upgrade speed, resulting in slow data access preview, update interruptions, and so on.

Dashboard Creation

In dashboard creation, the speed of display will be slowed down when there are too many levels or groups in a chart, or the number of components in one dashboard exceeds 30.

Optimizing the Server Performance

Optimizing Data Preparation

Data Process

• Process data properly. Try not to join tables N:N to avoid generating too many N:N relationships.

Data Update

• Avoid setting scheduling update for many single tables in different periods, and global update is recommended.

• Update tasks of the same time together to avoid repetitive updates of self-service dataset.

• Limit the frequency of updates by reducing updates when using the system or configure the resource pool parameters.

Data Usage

• Check database performance for direct connect data to ensure database access speed.

• Be careful about the usage of direct connect data.

Optimizing Dashboard

1. Data Type

If there is a large amount of data, switching direct connect data to extracted data is recommended.

If the speed of direct connect database is slow, you can optimize the SQL statements to reduce data.

2. Data Calculation

Check whether the dashboard has performed a large amount of calculation. For example, if there are distinct count, header filter, formula filter, and so on, you can reduce calculation, look for other calculation methods, or move the dashboard calculation to self-service dataset.

If you have requirements for chart big data grouping, see Upgrading Chart Big Data GCC.

3. Number of Components

Except for filter components, the number of components should not exceed 30.

The number of conditions of filter components should not exceed 30. One filter component may have multiple conditions.

For example, where City = "New York" or (City = "Los Angeles" and City = "Washington"). There may be one filter component on the front end, but it contains three filter conditions.