Hike News
Hike News

使用Arthas watch观察方法执行

java 本文总阅读量次

Arthas watch命令介绍

https://alibaba.github.io/arthas/watch.html

方便的观察到指定方法的调用情况。

能观察到的范围为：入参、返回值、抛出异常和当前对象的属性，通过编写 OGNL 表达式进行对应变量的查看。

使用方法

命令格式：

1
watch class method "{params,returnObj,throwExp}" -x 2

观察方法出参和返回值

1
2
3
4
5
6
7
8
9
10
11
12
13
14
$ watch demo.MathGame primeFactors "{params,returnObj}" -x 2
Press Ctrl+C to abort.
Affect(class-cnt:1 , method-cnt:1) cost in 44 ms.
ts=2018-12-03 19:16:51; [cost=1.280502ms] result=@ArrayList[
    @Object[][
        @Integer[535629513],
    ],
    @ArrayList[
        @Integer[3],
        @Integer[19],
        @Integer[191],
        @Integer[49199],
    ],
]

调整-x的值，观察具体的方法参数值

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
$ watch demo.MathGame primeFactors "{params,target}" -x 3
Press Ctrl+C to abort.
Affect(class-cnt:1 , method-cnt:1) cost in 58 ms.
ts=2018-12-03 19:34:19; [cost=0.587833ms] result=@ArrayList[
    @Object[][
        @Integer[47816758],
    ],
    @MathGame[
        random=@Random[
            serialVersionUID=@Long[3905348978240129619],
            seed=@AtomicLong[3133719055989],
            multiplier=@Long[25214903917],
            addend=@Long[11],
            mask=@Long[281474976710655],
            DOUBLE_UNIT=@Double[1.1102230246251565E-16],
            BadBound=@String[bound must be positive],
            BadRange=@String[bound must be greater than origin],
            BadSize=@String[size must be non-negative],
            seedUniquifier=@AtomicLong[-3282039941672302964],
            nextNextGaussian=@Double[0.0],
            haveNextNextGaussian=@Boolean[false],
            serialPersistentFields=@ObjectStreamField[][isEmpty=false;size=3],
            unsafe=@Unsafe[sun.misc.Unsafe@2eaa1027],
            seedOffset=@Long[24],
        ],
        illegalArgumentCount=@Integer[13159],
    ],
]

-x表示遍历深度，可以调整来打印具体的参数和结果内容，默认值是1。

条件表达式的例子

1
2
3
4
5
6
7
$ watch demo.MathGame primeFactors "{params[0],target}" "params[0]<0"
Press Ctrl+C to abort.
Affect(class-cnt:1 , method-cnt:1) cost in 68 ms.
ts=2018-12-03 19:36:04; [cost=0.530255ms] result=@ArrayList[
    @Integer[-18178089],
    @MathGame[demo.MathGame@41cf53f9],
]

按照耗时进行过滤

1
2
3
4
5
6
7
8
9
10
11
12
$ watch demo.MathGame primeFactors '{params, returnObj}' '#cost>200' -x 2
Press Ctrl+C to abort.
Affect(class-cnt:1 , method-cnt:1) cost in 66 ms.
ts=2018-12-03 19:40:28; [cost=2112.168897ms] result=@ArrayList[
    @Object[][
        @Integer[2141897465],
    ],
    @ArrayList[
        @Integer[5],
        @Integer[428379493],
    ],
]

#cost>200(单位是ms)表示只有当耗时大于200ms时才会输出，过滤掉执行时间小于200ms的调用

观察当前对象中的属性

1
2
3
4
5
6
7
$ watch demo.MathGame primeFactors 'target'
Press Ctrl+C to abort.
Affect(class-cnt:1 , method-cnt:1) cost in 52 ms.
ts=2018-12-03 19:41:52; [cost=0.477882ms] result=@MathGame[
    random=@Random[java.util.Random@522b408a],
    illegalArgumentCount=@Integer[13355],
]
java