Optimized APDEX calculation logic.
Excluded export requests and background FineDataLink tasks from APDEX calculations in Health Monitoring.
FineOps provides a Health Monitoring dashboard for intelligent monitoring of system health and operation status.
The Health Monitoring dashboard provides a user-experience-oriented monitoring platform that tracks system stability, performance, and O&M efficiency. It also identifies problematic requests and pinpoints their source objects (such as dashboards and templates).
The Health Monitoring function relies on the Tracing function. Ensure you have enabled Tracing and configured the corresponding global settings.
1. Log in to FineOps as the admin, select the O&M project, and choose Project Monitoring > Health Monitoring.
2. Select the required request type from the drop-down list as needed for filtering.
Platform operation requests, such as opening directories, searches, and permission calculations
Any non-data/non-resource request falls under this category.
3. Switch the monitoring time range as needed, where you can specify the time range after clicking Historical Analysis.
1. Indicator Data and Resources with High Memory Usage display the monitoring data in the last 24 hours.
2. Trend analysis panels (Health Status, User Usage Statistics, and Node Detail Table) and Exceptional Request List display the monitoring data in the selected time range. Time range options include the last 1/6/12/24/72 hours.
Panel Description
It displays five key indicators. The indicator with an abnormal value will be displayed in red.
Indicator Description
1. Indicator description: a comprehensive indicator indicating the system health
2. Calculation logic:
Comprehensive health score = (Satisfactory request count + Tolerable request count/2)/Total request count * 100%
Satisfactory requests: successful requests with a latency of less than 3 s
Tolerable requests: successful requests with a latency between 3 s (included) and 12 s (excluded)
3. Calculation scope:
Real-Time Monitoring: data in the last 24 hours
Historical Analysis: data in the selected month
4. Recommended value: above 95%
1. Indicator description: an industry standard used to evaluate application performance
FanRuan has elevated the scoring standards of APDEX, a critical indicator in FineOps, based on over a year of continuous observation and user research.
The APDEX calculation logic has been optimized in FineOps V2.3.0 with stricter performance thresholds and heightened sensitivity to performance fluctuations, enabling a more accurate reflection of user satisfaction in scenarios with issues such as authentication latency.
FanRuan will continuously optimize and improve the product, providing users with better performance experience.
APDEX = (Satisfactory request count + Tolerable request count/2)/Total request count * 100%
Satisfactory requests: data requests with a latency of less than 3 s, resource requests with a latency of less than 0.5 s, and configuration requests with a latency of less than 0.5 s
Tolerable requests: data requests with a latency between 3 s (included) and 12 s (excluded), resource requests with a latency between 0.5 s (included) and 2 s (excluded), and configuration requests with a latency between 0.5 s (included) and 2 s (excluded)
Export requests and background FineDataLink tasks are excluded from APDEX calculations in Health Monitoring.
2. Calculation logic: Request success rate = (Successful request count/Total request count) × 100%
Indicator description:
Real-Time Monitoring: the peak number of per-minute concurrent requests in the last 24 hours, calculated via request slicing
Historical Analysis: the peak number of per-minute concurrent requests in the selected month, calculated via request slicing
Real-Time Monitoring: the peak number of per-minute concurrent users in the last 24 hours, calculated via request slicing
Historical Analysis: the peak number of per-minute concurrent users in the selected month, calculated via request slicing
It displays performance indices and request success rates within the specified time range.
In Real-Time Monitoring mode, if you click a point and View Trace Detail, you will be redirected to the Trace tab page in Tracing. The view automatically displays spans occurring between one minute before and one minute after the selected time.
In Real-Time Monitoring mode, if you click a point and View Traffic Detail, you will be redirected to the Traffic Monitoring tab page. The view automatically displays traffic occurring between one minute before and one minute after the selected time.
In Real-Time Monitoring mode, if you click a point and Locate Current Time, all trend analysis panels show corresponding indicator data of the selected time.
Application Performance Index (Apdex)
3. Recommended value: above 95%
It displays the peak numbers of per-minute concurrent requests and users.
Page Description
It displays values of system health indicators within the specified time range for each node.
For details about the calculation logic of indicators, see the above content.
It shows dashboard performance within the specified time range.
The p90 curve indicates that 90% of requests complete faster than the value. The p95 and p99 curves follow the same logic.
In Real-Time Monitoring mode, if you click a point on the Service Response Time (ms) or Data Response Time (ms) chart and click View Trace Detail, you will be redirected to the Trace tab page in Tracing. The view automatically displays spans occurring between one minute before and one minute after the selected time.
In Real-Time Monitoring mode, if you click a point on the Blank Page Duration of User Application (ms) or Loading Duration of User Application (ms) chart and click View Trace Detail, you will be redirected to the First-Screen Trace tab page in Tracing. The view automatically displays time consumption details of traces occurring between one minute before and one minute after the selected time.
It displays the average latency of valid requests. It measures the time taken by data engines/databases to process requests, reflecting data computation performance.
It displays all error requests and slow requests (with a latency exceeding 10 seconds) within the specified time range.
The exceptional requests are classified by resource type. Error information of each exceptional request, including the exception type, the query count, and the affected user count, is displayed.
In Real-Time Monitoring mode, you can click View Trace Detail to jump to the Trace tab page in Tracing. The view automatically filters spans by session ID and displays spans occurring within the last three days. The filtering condition can be modified in Filter.
It displays resources with high memory usage identified within the last 24 hours or the selected month.
The resources are classified by resource type. Resource information, including the resource type, the resource name, the resource creator, the used memory, the access user, and the identification time, is displayed.
Administrators should conduct regular health inspections for applications to ensure proper configuration of the application environment and in-app items for normal application operation.
You are advised to inspect the system immediately and configure regular automatic inspections when the prompt appears: "The health inspection has not been conducted for a month. Start immediately for troubleshooting."
FineOps provides a Tracing function to help you collect and analyze requests.
You can click any abnormal point on charts in Health Monitoring and navigate to the corresponding traces to locate system performance issues.
In Real-Time Monitoring mode, you can click the View Trace Detail button in Exceptional Request List to view corresponding traces and locate system performance issues.