• Home
  • Line#
  • Scopes#
  • Navigate#
  • Raw
  • Download
1# Development of Performance Tracing
2
3## Introduction
4
5hiTraceMeter provides APIs for system performance tracing. You can call the APIs provided by the hiTraceMeter module in your own service logic to effectively track service processes and check the system performance.
6
7## Basic Concepts
8
9- **hiTraceMeter Tag**
10
11  Tag used for tracing data categorization. It is also known as **hiTraceMeter Category**. Generally, one subsystem maps to a tag. The tag is passed as the **Tag** parameter in performance tracing APIs. When you use the hiTraceMeter CLI tool to collect tracing data, only the tracing data specified by the **Tag** parameter is collected.
12
13## Working Principles
14
15- The application calls hiTraceMeter APIs to perform performance tracing. The APIs output the tracing data to the kernel's ftrace data buffer through the kernel's sysfs file interface.
16- The hiTraceMeter CLI tool reads the tracing data in the ftrace data buffer and saves the trace data as a text file on the device.
17
18## Constraints
19
20Due to the asynchronous I/O feature of JS, the hiTraceMeter module provides only asynchronous APIs.
21
22## Available APIs
23
24The performance tracing APIs are provided by the **hiTraceMeter** module. For details, see [API Reference]( ../reference/apis/js-apis-hitracemeter.md).
25
26**APIs for performance tracing**
27
28| API                                                                      | Return Value   | Description        |
29| ---------------------------------------------------------------------------- | --------- | ------------ |
30| hiTraceMeter.startTrace(name: string, taskId: number) | void      | Marks the start of a trace task. If multiple trace tasks with the same name need to be performed at the same time or a trace task needs to be performed multiple times concurrently, different task IDs must be specified in **startTrace**. If the trace tasks with the same name are not performed at the same time, the same task ID can be used.|
31| hiTraceMeter.finishTrace(name: string, taskId: number)                       | void      | Marks the end of a trace task. The values of **name** and **taskId** must be the same as those of **hiTraceMeter.startTrace**.|
32| hiTraceMeter.traceByValue(name: string, value: number)                       | void      | Marks the value changes of a numeric variable in a trace task.|
33
34## How to Develop
35
36In this example, distributed call chain tracing begins when the application startup execution page is loaded and stops when the service usage is completed.
37
381. Create a JS application project. In the displayed **Project** window, choose **entry** > **src** > **main** > **js** > **default** > **pages** > **index**, and double-click **index.js**. Add the code to implement performance tracing upon page loading. The sample code is as follows:
39
40   ```js
41   import hiTraceMeter from '@ohos.hiTraceMeter'
42
43   export default {
44       data: {
45           title: ""
46       },
47       onInit() {
48           this.title = this.$t('strings.world');
49
50           // Start trace tasks with the same name concurrently.
51           hiTraceMeter.startTrace("business", 1);
52           // Keep the service process running.
53           console.log(`business running`);
54           hiTraceMeter.startTrace("business", 2);  // Start the second trace task with the same name while the first task is still running. The tasks are running concurrently and therefore their taskId must be different.
55           // Keep the service process running.
56           console.log(`business running`);
57           hiTraceMeter.finishTrace("business", 1);
58           // Keep the service process running.
59           console.log(`business running`);
60           hiTraceMeter.finishTrace("business", 2);
61
62           // Start trace tasks with the same name in serial mode.
63           hiTraceMeter.startTrace("business", 1);
64           // Keep the service process running.
65           console.log(`business running`);
66           hiTraceMeter.finishTrace("business", 1);  // End the first trace task.
67           // Keep the service process running.
68           console.log(`business running`);
69           hiTraceMeter.startTrace("business", 1);   // Start the second trace task with the same name in serial mode.
70           // Keep the service process running.
71           console.log(`business running`);
72
73           let traceCount = 3;
74           hiTraceMeter.traceByValue("myTestCount", traceCount);
75           traceCount = 4;
76           hiTraceMeter.traceByValue("myTestCount", traceCount);
77           hiTraceMeter.finishTrace("business", 1);
78       }
79   }
80   ```
81
822. Create an ArkTs application project. In the displayed **Project** window, choose **entry** > **src** > **main** > **ets** > **pages** > **index**, and double-click **index.js**. Add the code to implement performance tracing upon page loading. For example, if the name of the trace task is **HITRACE\_TAG\_APP**, the sample code is as follows:
83
84   ```ts
85   import hitrace from '@ohos.hiTraceMeter';
86
87   @Entry
88   @Component
89   struct Index {
90     @State message: string = 'Hello World';
91
92     build() {
93       Row() {
94         Column() {
95           Text(this.message)
96             .fontSize(50)
97             .fontWeight(FontWeight.Bold)
98             .onClick(() => {
99               this.message = 'Hello ArkUI';
100
101               // Start trace tasks with the same name concurrently.
102               hitrace.startTrace("HITRACE_TAG_APP", 1001);
103               // Keep the service process running.
104               console.log(`HITRACE_TAG_APP running`);
105
106               // Start the second trace task with the same name while the first task is still running. The tasks are running concurrently and therefore their taskId must be different.
107               hitrace.startTrace("HITRACE_TAG_APP", 1002);
108               // Keep the service process running.
109               console.log(`HITRACE_TAG_APP running`);
110
111               hitrace.finishTrace("HITRACE_TAG_APP", 1001);
112               hitrace.finishTrace("HITRACE_TAG_APP", 1002);
113
114               // If trace tasks with the same name are not run concurrently, the same taskId can be used.
115               hitrace.startTrace("HITRACE_TAG_APP", 1003);
116               // Keep the service process running.
117               console.log(`HITRACE_TAG_APP running`);
118               // End the first trace task.
119               hitrace.finishTrace("HITRACE_TAG_APP", 1003);
120
121               // Start the second trace task with the same name in serial mode. It uses a taskId different from the first trace task.
122               hitrace.startTrace("HITRACE_TAG_APP", 1004);
123               // Keep the service process running.
124               console.log(`HITRACE_TAG_APP running`);
125               let traceCount = 3;
126               hitrace.traceByValue("myTestCount", traceCount);
127               hitrace.finishTrace("HITRACE_TAG_APP", 1004);
128
129               // Start the third trace task with the same name in serial mode. It uses a taskId same as the second trace task.
130               hitrace.startTrace("HITRACE_TAG_APP", 1004);
131               // Keep the service process running.
132               console.log(`HITRACE_TAG_APP running`);
133               // End the third trace task.
134               hitrace.finishTrace("HITRACE_TAG_APP", 1004);
135
136             })
137          }
138          .width('100%')
139        }
140        .height('100%')
141      }
142   }
143   ```
144
1453. Click the run button on the application page. Then, run the following commands in sequence in shell:
146
147   ```shell
148   hdc shell
149   hitrace --trace_begin app
150   ```
151
152   After the trace command is executed, call the hiTraceMeter APIs in your own service logic on the device. Then,  run the following commands in sequence:
153
154   ```shell
155   hitrace --trace_dump | grep tracing_mark_write
156   hitrace --trace_finish
157   ```
158
159   The following is an example of the captured trace data:
160
161   ```
162   <...>-3310    (-------) [005] .... 351382.921936: tracing_mark_write: S|3310|H:HITRACE_TAG_APP 1001
163   <...>-3310    (-------) [005] .... 351382.922138: tracing_mark_write: S|3310|H:HITRACE_TAG_APP 1002
164   <...>-3310    (-------) [005] .... 351382.922165: tracing_mark_write: F|3310|H:HITRACE_TAG_APP 1001
165   <...>-3310    (-------) [005] .... 351382.922175: tracing_mark_write: F|3310|H:HITRACE_TAG_APP 1002
166   <...>-3310    (-------) [005] .... 351382.922182: tracing_mark_write: S|3310|H:HITRACE_TAG_APP 1003
167   <...>-3310    (-------) [005] .... 351382.922203: tracing_mark_write: F|3310|H:HITRACE_TAG_APP 1003
168   <...>-3310    (-------) [005] .... 351382.922210: tracing_mark_write: S|3310|H:HITRACE_TAG_APP 1004
169   <...>-3310    (-------) [005] .... 351382.922233: tracing_mark_write: C|3310|H:myTestCount 3
170   <...>-3310    (-------) [005] .... 351382.922240: tracing_mark_write: F|3310|H:HITRACE_TAG_APP 1004
171   <...>-3310    (-------) [005] .... 351382.922247: tracing_mark_write: S|3310|H:HITRACE_TAG_APP 1004
172   <...>-3310    (-------) [005] .... 351382.922266: tracing_mark_write: F|3310|H:HITRACE_TAG_APP 1004
173   ```
174