FineBI server configuration recommendation

1.1 Version

1.2 Function introduction

FineBI is a pure B/S-side business intelligence analysis service platform; it supports deploying it on the server through the Web application server, and provides enterprise cloud server. The client only needs to use a browser to access and use the service platform.

FineBI uses the Spider engine, and the Spider data engine can flexibly support analysis of different data levels.

This article describes the recommended requirements for FineBI hardware and software configuration.

1.3 FineBI performance test report

The following direct and extract reports are only available for FineBI 5.1.5 and later.

Draw version performance test report: 

FineBI5.1 Test Report.pdf

Direct connection version performance test report: FineBI5.1.5 direct connection performance test report v1.0.pdf

Users need to refer to  the  real-time data & introduction of  extracted data to determine whether they are using "real-time data" or "extracted data".

2.1 All data are extracted data

When recommending the configuration, you need to consider the dimensions of "order of magnitude" and "number of users", and choose the higher configuration of the two.

Note: It is recommended to use "Solid State Drive" for a better experience.

2.1.1 Recommended configuration based on order of magnitude

When the maximum data volume of a single table is below 100 million, you can directly use the local disk of the Web server as the data storage medium. The recommended configuration is shown in the following table:

Note 1: JVM memory ≠ the whole machine memory, it is recommended that the JVM memory accounts for 2/3 ~ 3/4 of the whole machine memory.

Note 2: The available disk space here is the recommended space.

Note 3: The recommended disk read and write speed is at least 100M/S.

2.1.2 Recommended configuration according to the number of users

1) The recommended configuration is shown in the table below:

Note: JVM memory ≠ whole machine memory, it is recommended that JVM memory accounts for 2/3 ~ 3/4 of the whole machine memory.

2) The user type description is shown in the following table:

2.2 All data are real-time data

1) Performance and calculation all depend on the database, that is, real-time data. Due to the existence of a caching mechanism and memory calculation in some scenarios, the configuration of the Web server can be estimated by the amount of result set data. (The configuration of the user's database server is not recommended here.) Among them, the data volume of the result set represents the number of rows returned by the query data. The configuration recommendations are shown in the following table:

2) A lot of calculations here are done by the database, and BI memory consumption is mainly related to the result set size and concurrent performance.

• Editing users concurrently refers to the number of users who use FineBI to edit dashboards, create tables, and self-service datasets at the same time.

• Preview user concurrency refers to the number of users viewing data/dashboards using FineBI at the same time.

2.3 Both real-time data and extracted data

In the case of a mixture of both real-time and extracted data, the highest configuration requirement is sufficient.

The software environment used by FineBI is as follows:

Note 1: IE10 and below versions do not support full-screen viewing of templates. If IE11 and below versions enable Global Watermark , some operations may be affected, for example, components cannot be added to the dashboard.

Note 2: Google Chrome recommends using V70 and above.

3.1 Recommended installation environment

FineBI's web server should not be installed on a physical computer or VM virtual machine that simultaneously runs resource-intensive applications (such as databases or application servers). For details on virtual machines and physical machines, please refer to: Differences between virtual machines and physical machines .

In the CPU recommendation in the second section of this article, it is necessary to ensure that the actual resources that FineBI can occupy reach 80%.

3.2 Recommended Disks

FineBI is an IO-intensive application that relies heavily on disk IO, so it is recommended to use local disks or solid-state drives.

3.3 Recommended JVM memory accounts for 2/3 ~ 3/4 of the whole machine memory

JVM memory ≠ the whole machine memory, so it is not recommended to set the JVM memory to account for too much of the whole machine memory.

During BI running, in addition to JVM memory, it also needs to occupy out-of-heap memory (which can be configured by parameters), and also needs to reserve some memory for other applications such as db on the system and the machine to run.

Note: For details on configuring JVM memory, see: Deploying and Modifying Memory