如何在 Go 程序中使用 pprof?
有一个名为net/http/pprof的 Go 包,但我不能使用它。
该文件说go tool pprof http://localhost:6060/debug/pprof/heap
,这是行不通的。
而且,下面的_是什么意思?
import _ "net/http/pprof"
如何在 Go 程序中使用 pprof?
有一个名为net/http/pprof的 Go 包,但我不能使用它。
该文件说go tool pprof http://localhost:6060/debug/pprof/heap
,这是行不通的。
而且,下面的_是什么意思?
import _ "net/http/pprof"
根据您的评论,问题可能是您没有使用正确的端口号。
如果您正在运行 http 服务器http://localhost:9997
,那么我认为您希望使用以下命令运行命令http://localhost:9997
:
$ go tool pprof http://localhost:9997/debug/pprof/heap
根据net/http/pprof pkg doc page,如果您的应用程序已经在运行一个 http 服务器,您不需要启动一个,只需import _ "net/http/pprof"
在程序中包含某个地方。http://localhost:6060
是以服务器为例启动的,主机和端口是任意的。
import _ "net/http/pprof"
表示包已导入,但您不使用任何导出的标识符。根据go 语言规范,这将只为它的副作用导入包。我认为,这些副作用包括执行包源文件中定义的init() 函数,显然还有已注册的变量。
此外,您可能会发现这篇博文很有帮助:
“不起作用”是什么意思?什么是预期的,什么是观察的?
跑
$ go tool pprof -h
查看此工具版本的帮助。我的本地版本在发布时(即介于两者之间)都没有提示。表明:
(10:16) jnml@fsc-r550:~$ go tool pprof -h
Option h is ambiguous (heapcheck, help)
Invalid option(s)
Usage:
pprof [options] <program> <profiles>
<profiles> is a space separated list of profile names.
pprof [options] <symbolized-profiles>
<symbolized-profiles> is a list of profile files where each file contains
the necessary symbol mappings as well as profile data (likely generated
with --raw).
pprof [options] <profile>
<profile> is a remote form. Symbols are obtained from host:port/pprof/symbol
Each name can be:
/path/to/profile - a path to a profile file
host:port[/<service>] - a location of a service to get profile from
The /<service> can be /pprof/heap, /pprof/profile, /pprof/pmuprofile,
/pprof/growth, /pprof/contention, /pprof/wall,
/pprof/thread, or /pprof/filteredprofile.
For instance:
pprof http://myserver.com:80/pprof/heap
If /<service> is omitted, the service defaults to /pprof/profile (cpu profiling).
pprof --symbols <program>
Maps addresses to symbol names. In this mode, stdin should be a
list of library mappings, in the same format as is found in the heap-
and cpu-profile files (this loosely matches that of /proc/self/maps
on linux), followed by a list of hex addresses to map, one per line.
For more help with querying remote servers, including how to add the
necessary server-side support code, see this filename (or one like it):
/usr/doc/google-perftools-1.5/pprof_remote_servers.html
Options:
--cum Sort by cumulative data
--base=<base> Subtract <base> from <profile> before display
--interactive Run in interactive mode (interactive "help" gives help) [default]
--seconds=<n> Length of time for dynamic profiles [default=30 secs]
--add_lib=<file> Read additional symbols and line info from the given library
--lib_prefix=<dir> Comma separated list of library path prefixes
Reporting Granularity:
--addresses Report at address level
--lines Report at source line level
--functions Report at function level [default]
--files Report at source file level
Output type:
--text Generate text report
--callgrind Generate callgrind format to stdout
--gv Generate Postscript and display
--web Generate SVG and display
--list=<regexp> Generate source listing of matching routines
--disasm=<regexp> Generate disassembly of matching routines
--symbols Print demangled symbol names found at given addresses
--dot Generate DOT file to stdout
--ps Generate Postcript to stdout
--pdf Generate PDF to stdout
--svg Generate SVG to stdout
--gif Generate GIF to stdout
--raw Generate symbolized pprof data (useful with remote fetch)
Heap-Profile Options:
--inuse_space Display in-use (mega)bytes [default]
--inuse_objects Display in-use objects
--alloc_space Display allocated (mega)bytes
--alloc_objects Display allocated objects
--show_bytes Display space in bytes
--drop_negative Ignore negative differences
Contention-profile options:
--total_delay Display total delay at each region [default]
--contentions Display number of delays at each region
--mean_delay Display mean delay at each region
Call-graph Options:
--nodecount=<n> Show at most so many nodes [default=80]
--nodefraction=<f> Hide nodes below <f>*total [default=.005]
--edgefraction=<f> Hide edges below <f>*total [default=.001]
--focus=<regexp> Focus on nodes matching <regexp>
--ignore=<regexp> Ignore nodes matching <regexp>
--scale=<n> Set GV scaling [default=0]
--heapcheck Make nodes with non-0 object counts
(i.e. direct leak generators) more visible
Miscellaneous:
--tools=<prefix> Prefix for object tool pathnames
--test Run unit tests
--help This message
--version Version information
Environment Variables:
PPROF_TMPDIR Profiles directory. Defaults to $HOME/pprof
PPROF_TOOLS Prefix for object tools pathnames
Examples:
pprof /bin/ls ls.prof
Enters "interactive" mode
pprof --text /bin/ls ls.prof
Outputs one line per procedure
pprof --web /bin/ls ls.prof
Displays annotated call-graph in web browser
pprof --gv /bin/ls ls.prof
Displays annotated call-graph via 'gv'
pprof --gv --focus=Mutex /bin/ls ls.prof
Restricts to code paths including a .*Mutex.* entry
pprof --gv --focus=Mutex --ignore=string /bin/ls ls.prof
Code paths including Mutex but not string
pprof --list=getdir /bin/ls ls.prof
(Per-line) annotated source listing for getdir()
pprof --disasm=getdir /bin/ls ls.prof
(Per-PC) annotated disassembly for getdir()
pprof http://localhost:1234/
Enters "interactive" mode
pprof --text localhost:1234
Outputs one line per procedure for localhost:1234
pprof --raw localhost:1234 > ./local.raw
pprof --text ./local.raw
Fetches a remote profile for later analysis and then
analyzes it in text mode.
FATAL ERROR: Invalid option(s)
go tool pprof: exit status 1
(10:16) jnml@fsc-r550:~$
关于第二个问题: `import _ "foo"' 习语是导入包 foo 只是为了 foo 初始化的副作用。这可能包括例如注册由 foo 提供的功能,以便在其他包中可用。一个具体的例子是特定的图像格式处理包(见底部的“子目录”)和抽象图像包。