You (the admin) can help users collect and analyze the trace details in the query request in case of dashboard/report lagging, without passively waiting for user feedback.
It allows for filtering requests by ID, time, and other conditions, to facilitate further analysis and pinpointing of slow requests and precise identification of issues such as network fluctuations, database performance, abnormal configuration database, complex resources, and resource shortages.
You have enabled the Tracing function and configured related parameters in global settings before using this function.
For details, see Prerequisites for Using Tracing.
1. Log in to FineOps as the admin and select the required O&M project. Choose Tracing > Trace.
2. Select a project component in the upper right corner. Tracing is supported for FineReport, FineBI - Application Node, Engine - Metadata Node, Engine - Calculation Node, and Nginx components of projects only.
3. Click Filter, configure the query conditions, and click Search.
The following table describes the supported query conditions.
Description: Searches for all queries started within the specified period.
Option: any time range within the past five days from the current date, but the span cannot exceed three days
Default: the past 30 minutes from the current time
Description: Searches for records related to the specified session.
Option: IDs of all sessions within the selected time range
Default: empty, meaning show all
Description: Searches for traces by ID.
Option: IDs of all traces within the set time range
Description: Searches for all queries triggered by the specified user.
Option: all recorded users
Description: Searches for query records corresponding to the accessed object.
Option:
FineReport project: actual template path
FineBI project: dashboard name and component name
Description: Searches for query records where the loading time of traces falls within the specified range.
Option: non-negative integers (If it is left empty, no upper/lower limit will be set.)
Description: Filters traces by type.
Option: all endpoints within the set time range
Description: Filters traces by the success/failure status of the request.
Success: Obtains traces from successful requests.
Failure: Obtains traces from failed requests.
Description: Filters traces by node within the O&M project.
Option: each cluster node
Description: Searches for queries where all tags match the selected values.
Option: sessionId, reportId, dec.user.name, fr.domain, fr.cmd, db.type, http.method, http.status_code, and db.instance (The last five options cannot be selected simultaneously.)
The page after you click Search is shown in the following figure.
You can sort the query result traces:
Ascend by Duration: Sorts queries by their first-screen loading time from shortest to longest.
Descend by Duration: Sorts queries by their first-screen loading time from longest to shortest.
Ascend by Start Time: Sorts queries by their trigger time from oldest to most recent.
Descend by Start Time: Sorts queries by their trigger time from most recent to oldest.
Internal users can click this button to export trace data, which can be provided to FanRuan technical support personnel for troubleshooting.
The first 100 trace data records after sorting will be exported.
It displays summary information of each trace, including:
Accessed content: It consists of the request type (POST or GET) and the URL.
Trace duration: The duration is highlighted in different colors, where yellow indicates requests exceeding 10 s, red indicates request failure, and grey indicates others.
FGC tag: It indicates that a Full GC occurred during the trace timeframe.
Time: The timestamp when the trace occurred is displayed.
You can click this button to collapse or expand the left-hand trace list panel.
1. List/Analysis Summary
The List/Analysis Summary tab page displays details of each span in a corresponding view. You can click tabs to toggle between views.
2. Trace ID
The trace ID is displayed in the upper right corner. You can click the copy icon to copy the trace ID.
3. Get Log
Prerequisite: The O&M project must use the Elasticsearch suite for log collection and storage.
Instruction: You can click the Get Log button to download O&M logs of the trace with the corresponding ID to local.
4. View First-Screen Trace
If a session associated with the current trace has relevant first-screen trace data, you can click this button to view details on the First-Screen Trace page.
1. Trace ID copy icon: You can click this icon to quickly copy the trace ID for log queries.
2. Get Log button
3. Duration filter: You can limit the duration to search for traces that fall within the range.
4. Span keyword search: You can search for spans whose names contain the specified keywords.
5. The span details of the trace are displayed in the form of a detail table. Clicking a span reveals more details.
6. You can click the Trace Information button to view the most critical trace information extracted by the system, without manually inspecting each span.
The Analysis Summary tab page consists of two sections.
1. Duration distribution: This section helps understand the breakdown of time consumption at a glance, which is categorized into the duration by peer data fetching and others.
Clicking a specific category will dynamically filter the spans listed in the span summary section below.
2. Span summary: This section displays a summary of span information for the entire trace.
The percentage of the total trace duration consumed by this span
(Content in this table is sorted in descending order by this column.)
The category of the time consumption, which is the specific peer instance responsible for data fetching, or Others
3. Clicking a span name will redirect you to the List tab page, where the search is automatically performed with that name as the keyword.