1. Allowed selecting Median as Statistical Method of performance data to reduce the impact of abnormal data.
2. Added a distribution chart on the individual template performance page to help you quickly understand the loading time of the template during each access.
3. Optimized field names in the detail table for better understanding.
You serve as an administrator responsible for managing FanRuan applications. While projects operate within acceptable parameters, you have encountered persistent performance-related issues reported by business users, specifically concerning template functionality. These issues manifest as preview freezing, prolonged loading time, and noticeable latency during operation.
You need to identify the root causes of slow-performing templates by addressing:
Which templates exhibit slow performance?
At which stage does the slowdown occur?
Which functions contribute to the performance degradation?
What data queries are causing the delays?
Only by pinpointing the underlying causes of these performance issues can you implement targeted optimizations to enhance user experience and mitigate risks caused by excessive system loads.
FineOps provides a Template Performance function, which can help you (the admin) improve:
1. Development efficiency: This feature enhances the efficiency of template performance detection and troubleshooting, thereby accelerating the development of business personnel.
2. User satisfaction: By improving system stability and optimizing the end-user experience, this function contributes to higher levels of user satisfaction.
3. Cost-effectiveness: This function identifies and addresses performance issues at their source and offers comprehensive diagnostic insights, reducing the performance maintenance costs for admins.
1. This function is supported exclusively for FineReport general reports, FineReport dashboards, and FineBI dashboards opened on the PC desktop.
2. You have enabled the Tracing function and configured global settings before using this function.
For details, see Prerequisites for Using Tracing.
3. You have configured Slow Request Standard in Tracing. Only requests with a response time exceeding the set threshold are classified as slow requests.
Log in to FineOps as the admin, choose Setting > Tracing, set Slow Request Standard, and click Save.
1. Log in to FineOps as the admin and select the required O&M project. Choose Health Inspection > Template Performance.
2. Select Statistical Method.
You can specify the calculation method for performance metrics including Loading Time, Number of Slow Requests, and Data Fetching Time.
You can choose between Median (default selection) and Average.
3. (Optional) Tick Latest Version Only.
If it is unticked, the system displays performance data of both the current and historical versions of the template.
If it is ticked, the system displays performance data of the latest version of the template only.
4. Set the filtering condition.
Click Filter and configure the filtering condition. You can filter templates based on multiple dimensions. Click Search.
You can set the start time and the end time.
The start time defaults to five days prior to the current time.
The end time defaults to six hours after the current time.
The default option is All.
You can choose between Mounted on Platform and Not Mounted on Platform.
All templates are selected by default. Multi-selection is supported.
You can select any template in the chosen mounting status.
You can filter templates with first-page loading time exceeding the set duration.
It is empty by default and allows the input of positive integers.
All templates with slow performance that meet the filtering condition are displayed in a detail table.
The following table describes the detail table content.
For a FineReport report, it refers to the full path of the template in the reportlet directory of the project.
For a FineBI dashboard, it refers to the full path of the dashboard in the user's analysis.
It refers to the average or median value of the first-page loading time.
The first-page loading time refers to the time to load the visible content in a template for interaction after the template is opened.
If multiple FineBI dashboards are opened simultaneously, the loading time of some dashboards may not be included in the summary calculation.
It refers to the number of requests with a response time exceeding the threshold defined in Slow Request Standard.
The data fetching time of a single request (X) is defined as the total time spent fetching data from all datasets associated with the request.
The data fetching time of a single access (Y) is defined as the average or median value of X across all requests during that access.
The data fetching time of a template is defined as the average or median value of Y across all accesses of that template.
It refers to the data fetching type of FineBI dashboards only, including Data Extraction and Direct Connection.
You can click the Template Detail button of a template on the overall template performance page to be redirected to the performance page of the template.
You can adjust the template while clicking the Refresh icon in the upper right corner to view the real-time effect.
The performance detail page of a single template consists of a distribution chart and a detail table.
1. Distribution chart:
Each dot represents an access, allowing you to view the template loading time of each access.
The dashed line in the chart indicates the average or median value of the template loading time.
2. Detail table:
It refers to the version of the accessed template and the value is the last modification time of the template configuration.
If multiple templates are opened simultaneously, the loading time may be skewed and is not referential.
You can be redirected to the chain detail page to view the chain of requests at specific time.
If the time is beyond the chain retention period, the chain details cannot be viewed.
For details, see First-Page Chain and Call Chain.
The following table describes performance risks.
The network transmission takes a long time.
The DNS resolution takes a long time.
Fetching data from the dataset is slow.
Longer than 3 s
The volume of data fetched by a widget is large.
The data volume is greater than 10,000.
The cell sorting calculation takes a long time.
The cell conditional attribute calculation takes a long time.
The cell display calculation takes a long time.
The cell expansion calculation happens too many times.
The collapsible tree construction takes a long time.
The chart formula calculation takes a long time.
The data extraction is slow.
The data fetching based on direct connections is slow.
The permission query takes a long time.