Flask-OpenTracing 2.0.0

初始化

FlaskTracing包装了opentracing支持的tracer实例。要创建一个FlaskTracing对象，您可以传入tracer对象直接或是一个返回tracer对象的调用。例如

import opentracing
from flask_opentracing import FlaskTracing

opentracing_tracer = ## some OpenTracing tracer implementation
tracing = FlaskTracing(opentracing_tracer, ...)

或

import opentracing
from flask_opentracing import FlaskTracing

def initialize_tracer():
    ...
    return opentracing_tracer

tracing = FlaskTracing(initialize_tracer, ...)

跟踪所有请求

import opentracing
from flask_opentracing import FlaskTracing

app = Flask(__name__)

opentracing_tracer = ## some OpenTracing tracer implementation
tracing = FlaskTracing(opentracing_tracer, True, app, [optional_args])

跟踪单个请求

import opentracing
from flask_opentracing import FlaskTracing

app = Flask(__name__)

opentracing_tracer = ## some OpenTracing tracer implementation
tracing = FlaskTracing(opentracing_tracer)

@app.route('/some_url')
@tracing.trace(optional_args)
def some_view_func():
    ...
    return some_view

手动访问span

为了访问请求的span，我们提供了一个方法FlaskTracing.get_span(request)，它返回请求的span，如果它存在且未完成。这可以用于将重要事件记录到span、设置标签或创建子span以跟踪非RPC事件。如果没有传递请求，则将使用当前请求。

跟踪RPC

如果您想进行RPC并继续现有的跟踪，可以将当前span注入到RPC中。例如，如果进行HTTP请求，以下代码将在网络中继续您的跟踪。

@tracing.trace()
def some_view_func(request):
    new_request = some_http_request
    current_span = tracing.get_span(request)
    text_carrier = {}
    opentracing_tracer.inject(span, opentracing.Format.TEXT_MAP, text_carrier)
    for k, v in text_carrier.iteritems():
        new_request.add_header(k,v)
    ... # make request