skywalking 通过python探针监控Python 微服务应用性能

来源：https://skywalking.apache.org/zh/2020-11-30-pycon/

Python 作为一门功能强大的编程语言，被广泛的应用于计算机行业之中； 在微服务系统架构盛行的今天，Python 以其丰富的软件生态和灵活的语言特性在服务端编程领域也占有重要的一席之地。 本次分享将阐述 Apache SkyWalking 在微服务架构中要解决的问题，展示如何使用 Apache SkyWalking 来近乎自动化地监控 Python 后端应用服务，并对 Apache SkyWalking 的 Python 语言探针的实现技术进行解读。

python探针源码：https://github.com/apache/skywalking-python/tree/v0.6.0#create-spans

SkyWalking Python Agent

SkyWalking-Python: The Python Agent for Apache SkyWalking, which provides the native tracing abilities for Python project.

SkyWalking: an APM(application performance monitor) system, especially designed for microservices, cloud native and container-based (Docker, Kubernetes, Mesos) architectures.

 

Install

From Pypi

The Python agent module is published to Pypi, from where you can use pip to install:

# Install the latest version, using the default gRPC protocol to report data to OAP
pip install "apache-skywalking"

# Install the latest version, using the http protocol to report data to OAP
pip install "apache-skywalking[http]"

# Install the latest version, using the kafka protocol to report data to OAP
pip install "apache-skywalking[kafka]"

# Install a specific version x.y.z
# pip install apache-skywalking==x.y.z
pip install apache-skywalking==0.1.0  # For example, install version 0.1.0 no matter what the latest version is

From Source Codes

Refer to the FAQ.

Set up Python Agent

SkyWalking Python SDK requires SkyWalking 8.0+ and Python 3.5+.

If you want to try out the latest features that are not released yet, please refer to the guide to build from sources.

from skywalking import agent, config

config.init(collector='127.0.0.1:11800', service='your awesome service')
agent.start()

Alternatively, you can also pass the configurations via environment variables (such as SW_AGENT_NAME, SW_AGENT_COLLECTOR_BACKEND_SERVICES, etc.) so that you don't need to call config.init.

All supported environment variables can be found here

Supported Libraries

There are some built-in plugins (such as http.server, Flask, Django etc.) that support automatic instrumentation of Python libraries, the complete lists can be found here

API

Apart from the libraries that can be instrumented automatically, we also provide some APIs to enable manual instrumentation.

Create Spans

The code snippet below shows how to create entry span, exit span and local span.

from skywalking import Component
from skywalking.trace.context import SpanContext, get_context
from skywalking.trace.tags import Tag

context: SpanContext = get_context()  # get a tracing context
# create an entry span, by using `with` statement,
# the span automatically starts/stops when entering/exiting the context
with context.new_entry_span(op='https://github.com/apache') as span:
    span.component = Component.Flask
# the span automatically stops when exiting the `with` context

with context.new_exit_span(op='https://github.com/apache', peer='localhost:8080') as span:
    span.component = Component.Flask

with context.new_local_span(op='https://github.com/apache') as span:
    span.tag(Tag(key='Singer', val='Nakajima'))

Decorators

from time import sleep

from skywalking import Component
from skywalking.decorators import trace, runnable
from skywalking.trace.context import SpanContext, get_context
from skywalking.trace.ipc.process import SwProcess

@trace()  # the operation name is the method name('some_other_method') by default
def some_other_method():
    sleep(1)


@trace(op='awesome')  # customize the operation name to 'awesome'
def some_method():
    some_other_method()


@trace(op='async_functions_are_also_supported')
async def async_func():
    return 'asynchronous'


@trace()
async def async_func2():
    return await async_func()


@runnable() # cross thread propagation
def some_method(): 
    some_other_method()

from threading import Thread 
t = Thread(target=some_method)
t.start()

# When another process is started, agents will also be started in other processes, 
# supporting only the process mode of spawn.
p1 = SwProcess(target=some_method) 
p1.start()
p1.join()


context: SpanContext = get_context()
with context.new_entry_span(op=str('https://github.com/apache/skywalking')) as span:
    span.component = Component.Flask
    some_method()

Contact Us

• Submit an issue by using [Python] as title prefix.
• Mail list: dev@skywalking.apache.org. Mail to dev-subscribe@skywalking.apache.org, follow the reply to subscribe the mail list.
• Join skywalking channel at Apache Slack. If the link is not working, find the latest one at Apache INFRA WIKI.
• Twitter, ASFSkyWalking

Contributing

Before submitting a pull request or push a commit, please read our contributing and developer guide.

FAQs

Check the FAQ page or add the FAQs there.

License

Apache 2.0