python-eureka-client

Discription

This is an eureka client written in python, you can easily intergrate your python components with spring cloud.

Support Python Version

Python 3.7+

From 0.9.0, python 2 is no longer supported, if you are using python 2, please use version 0.8.12.

Why choose

• Register your python components to eureka server.
• Support failover.
• Support DNS discovery.
• Send heartbeat to eureka server.
• Pull registry from eureka server.
• Easy to use interface to use other REST service.
• HA when calling other REST service.
• Both trandictional and async def interfaces are provided.
• The http client lib is replacable.

How to use

Install

pip install py_eureka_client

Getting Start

This is the easiest way to use this component.

import py_eureka_client.eureka_client as eureka_client

your_rest_server_port = 9090
# The flowing code will register your server to eureka server and also start to send heartbeat every 30 seconds
eureka_client.init(eureka_server="http://your-eureka-server-peer1,http://your-eureka-server-peer2",
                   app_name="your_app_name",
                   instance_port=your_rest_server_port)

Then, in your business code, use

import py_eureka_client.eureka_client as eureka_client

res = eureka_client.do_service("OTHER-SERVICE-NAME", "/service/context/path")
print("result of other service" + res)

You can also use the EurekaClient class.

from py_eureka_client.eureka_client import EurekaClient
client = EurekaClient(eureka_server="http://my_eureka_server_peer_1/eureka/v2,http://my_eureka_server_peer_2/eureka/v2", app_name="python_module_1", instance_port=9090)
await client.start()
res = await client.do_service("OTHER-SERVICE-NAME", "/service/context/path")
print("result of other service" + res)
# when server is shutted down:
await client.stop()

In fact, the init function is a facade of the EurekaClient, it holds a client object behind, you can get that by catching its return value or use eureka_client.get_client() to get it. The init function will automatically start the client, while using raw EurekaClient, you must call the start() and stop() method explicitly.

From 0.11.0, all the methods in EurekaClient are defined async, and there are also async facade for init, do_servise, stop functions names init_async, do_service_async, sto_async.

Registering to Eureka Server

The most common method to will be like:

import py_eureka_client.eureka_client as eureka_client

your_rest_server_port = 9090
eureka_client.init(eureka_server="http://your-eureka-server-peer1,http://your-eureka-server-peer2",
                                app_name="python_module_1",
                                instance_port=your_rest_server_port)

But if you have deploy your eureka server in several zones, you should specify the eureka_availability_zones parameter.

import py_eureka_client.eureka_client as eureka_client
eureka_client.init(eureka_availability_zones={
                "us-east-1c": "http://ec2-552-627-568-165.compute-1.amazonaws.com:7001/eureka/v2/,http://ec2-368-101-182-134.compute-1.amazonaws.com:7001/eureka/v2/",
                "us-east-1d": "http://ec2-552-627-568-170.compute-1.amazonaws.com:7001/eureka/v2/",
                "us-east-1e": "http://ec2-500-179-285-592.compute-1.amazonaws.com:7001/eureka/v2/"}, 
                zone="us-east-1c",
                app_name="python_module_1", 
                instance_port=9090,
                data_center_name="Amazon")

If you are looking for flexibility, you should configure Eureka service URLs using DNS.

For instance, following is a DNS TXT record created in the DNS server that lists the set of available DNS names for a zone.

txt.us-east-1.mydomaintest.netflix.net="us-east-1c.mydomaintest.netflix.net" "us-east-1d.mydomaintest.netflix.net" "us-east-1e.mydomaintest.netflix.net"

Then, you can define TXT records recursively for each zone similar to the following (if more than one hostname per zone, space delimit)

txt.us-east-1c.mydomaintest.netflix.net="ec2-552-627-568-165.compute-1.amazonaws.com" "ec2-368-101-182-134.compute-1.amazonaws.com"
txt.us-east-1d.mydomaintest.netflix.net="ec2-552-627-568-170.compute-1.amazonaws.com"
txt.us-east-1e.mydomaintest.netflix.net="ec2-500-179-285-592.compute-1.amazonaws.com"

And then you can create the client like:

import py_eureka_client.eureka_client as eureka_client
eureka_client.init(eureka_domain="mydomaintest.netflix.net",
                region="us-east-1",
                zone="us-east-1c",
                app_name="python_module_1", 
                instance_port=9090,
                data_center_name="Amazon")

You can specify the protocol, basic authentication and context path of your eureka server separatly rather than setting it at the URL.

import py_eureka_client.eureka_client as eureka_client
eureka_client.init(eureka_domain="mydomaintest.netflix.net",
                region="us-east-1",
                zone="us-east-1c",
                eureka_protocol="https",
                eureka_basic_auth_user="keijack",
                eureka_basic_auth_password="kjauthpass",
                eureka_context="/eureka/v2",
                app_name="python_module_1", 
                instance_port=9090,

or

import py_eureka_client.eureka_client as eureka_client
eureka_client.init(eureka_server="your-eureka-server-peer1,your-eureka-server-peer2",
                eureka_protocol="https",
                eureka_basic_auth_user="keijack",
                eureka_basic_auth_password="kjauthpass",
                eureka_context="/eureka/v2",
                app_name="python_module_1", 
                instance_port=9090)

About the instance IP and hostname:

If you are using a Amazon data center, py-eureka-client will try to use local-ipv4 and local-hostname get from Amazon metadata service. In other cases, py-eureka-client will use the first non-loopback ip address and hostname from your net interface.

You can also specify both these tow field or just one of them explicitly:

eureka_client.init(eureka_server="your-eureka-server-peer1,your-eureka-server-peer2",
                eureka_protocol="https",
                eureka_basic_auth_user="keijack",
                eureka_basic_auth_password="kjauthpass",
                eureka_context="/eureka/v2",
                app_name="python_module_1", 
                instance_ip="192.168.10.168",
                instance_host="my-py-component.mydomian.com",
                instance_port=9090)

In some case you might have more than one interfaces attached, for example, you are running your application in a docker-container. In this case you can specify a network via instance_ip_network to be used to get the container's ip and host. You can use:

eureka_client.init(eureka_server="your-eureka-server-peer1,your-eureka-server-peer2",
                eureka_protocol="https",
                eureka_basic_auth_user="keijack",
                eureka_basic_auth_password="kjauthpass",
                eureka_context="/eureka/v2",
                app_name="python_module_1", 
                instance_ip_network="192.168.10.0/24",
                instance_port=9090)

If you want to get the ip only and sepecify the host by yourself, try:

import py_eureka_client.netint_utils as netint_utils

# you can get the ip only
ip = netint_utils.get_first_non_loopback_ip("192.168.10.0/24")
host = "my-py-component.mydomian.com"

eureka_client.init(eureka_server="your-eureka-server-peer1,your-eureka-server-peer2",
                eureka_protocol="https",
                eureka_basic_auth_user="keijack",
                eureka_basic_auth_password="kjauthpass",
                eureka_context="/eureka/v2",
                app_name="python_module_1", 
                instance_ip=ip,
                instance_host=host,
                instance_port=9090)

Error Callback

You can specify a callback function when initializing the eureka client, when errors occur in register, discover or status update phase, the callback function will be called to inform you. The callback function will be called only when all the eureka server url are all tried and fails.

The callback function should accept 2 arguments. which are the error type and the exception itself. please check:

def on_err(err_type: str, err: Exception):
    if err_type in (eureka_client.ERROR_REGISTER, eureka_client.ERROR_DISCOVER):
        eureka_client.stop()
    else:
        print(f"{err_type}::{err}")

your_rest_server_port = 9090
eureka_client.init(eureka_server="http://your-eureka-server-peer1,http://your-eureka-server-peer2",
                                app_name="python_module_1",
                                instance_port=your_rest_server_port,
                                on_error=on_err)

Call Remote Service

After init the eureka client, this is the most simplist way to do service:

import py_eureka_client.eureka_client as eureka_client

try:
    res = eureka_client.do_service("OTHER-SERVICE-NAME", "/service/context/path")
    print("result of other service" + res)
except urllib.request.HTTPError as e:
    # If all nodes are down, a `HTTPError` will raise.
    print(e)

do_service function also recieve a return_type keyword parameter, which when json was passed, the result will be a dict type object whereas response_object is pass, the original HTTPResponse object will be return. Please read the relative document for more information.

You can also use its async version:

import py_eureka_client.eureka_client as eureka_client

res = await eureka_client.do_service_async("OTHER-SERVICE-NAME", "/service/context/path")

do_service method will automatically try other nodes when one node return a HTTP error, until one success or all nodes being tried.

If you want to handle all the services' calling, you can use walk_nodes function:

import py_eureka_client.eureka_client as eureka_client

# you can define this function with `async def`
def walk_using_your_own_urllib(url):
    print(url)
    """
    # Connect to url and read result, then return it.
    # The result you return here will be returned to the `eureka_client.walk_nodes` function
    # If you want find this node is down, you can raise a `urllib.request.HTTPError`(urllib2.HTTPError in python2)
    # Then the `eureka_client.walk_nodes` will try to find another node to do the service.
    """

# result is the result that you return in walk_using_your_own_urllib function
try:
    res = eureka_client.walk_nodes("OTHER-SERVICE-NAME", "/service/context/path", walker=walk_using_your_own_urllib)
    print(res)
except urllib.request.HTTPError as e:
    # If all nodes are down, a `HTTPError` will raise.
    print(e)

A async version is also provied:

import py_eureka_client.eureka_client as eureka_client

def walk_using_your_own_urllib(url):
    print(url)
    """
    # Connect to url and read result, then return it.
    # The result you return here will be returned to the `eureka_client.walk_nodes` function
    # If provided node is down, you can raise a `urllib.request.HTTPError`(urllib2.HTTPError in python2)
    # Then the `eureka_client.walk_nodes` will try to find another node to do the service.
    """

res = await eureka_client.walk_nodes_async("OTHER-SERVICE-NAME", "/service/context/path",
                          walker=walk_using_your_own_urllib)

High Available Strategies

There are several HA strategies when using discovery client. They are:

• HA_STRATEGY_RANDOM, default strategy, find an node randamly.
• HA_STRATEGY_STICK, use one node until it goes down.
• HA_STRATEGY_OTHER, always use a different node from the last time.

In your init function, you can specify one of the above strategies:

import py_eureka_client.eureka_client as eureka_client
# General init method
eureka_client.init(eureka_server="http://your-eureka-server-peer1,http://your-eureka-server-peer2",
                   app_name="your_app_name",
                   instance_port=your_rest_server_port,
                   ha_strategy=eureka_client.HA_STRATEGY_STICK)

If the build-in stratergies do not satify you, you can load all the registry by following code:

import py_eureka_client.eureka_client as eureka_client

client = eureka_client.get_client()
app = client.applications.get_application("OTHER-SERVICE-NAME")
up_instances = app.up_instances
up_instances_same_zone = app.up_instances_in_zone(client.zone)
up_instances_other_zone = app.up_instances_not_in_zone(client.zone)
inst = up_instances[0]

# ... construct your url and do the service call

Use Other Http Client

You can use other http client to connect to eureka server and other service rather than the build-in urlopen method. It should be useful if you use https connections via self-signed cetificates.

From 0.11.0, the methods of the http_client.HttpClient are defined async, you can now use some async http libs like aiohttp

To do this, you should:

1. (Optional) At most scenario, you should also write a class that inherited from py_eureka_client.http_client.HttpResponse, for the reason of the py_eureka_client.http_client.HttpResponse class wraps the http.client.HTTPResponse which may not return by the third http libs.
2. Write a class inherited the HttpClient class in py_eureka_client.http_client.
3. Rewrite the urlopen method in your class. this method must return an subclass of py_eureka_client.http_client.HttpResponse, which is a wrapper class that hold to properties called raw_response and body_text.
4. Set you own HttpClient object into py_eureka_client.http_client by py_eureka_client.set_http_client

import py_eureka_client.http_client as http_client

# 1. A class inherited `py_eureka_client.http_client.HttpResonse`

class MyHttpResponse(http_client.HttpResponse):

    def __init__(self, raw_response):
        """
        " This raw response will return when you pass `response_object` in the `do_service` function.
        """
        self.raw_response = raw_response
    
    @property
    def body_text(self):
        txt = ""
        """
        " Read the body text from `self.raw_response`
        """
        return txt

# 2. A class inherited `py_eureka_client.http_client.HttpClient`.
class MyHttpClient(http_client.HttpClient):

    # 3. Rewrite the `urlopen` method in your class.
    # If you want to raise an exception, please make sure that the exception is an `http_client.HTTPError` or `http_client.URLError`.     
    async def urlopen(self, request: Union[str, http_client.HttpRequest] = None,
                      data: bytes = None, timeout: float = None) -> http_client.HttpResponse:
        res = await your_own_http_client_lib.do_the_visit(request, data, timeout)
        return MyHttpResponse(res)
        # You can parse your response object here, and set the body_text to http_client.HttpResponse, then you may ignore the http_client.HttpResponse inheritance.
        # body_txt = parse_res_body(res)
        # http_res = http_client.HttpResponse()
        # http_res.raw_response = res
        # http_res.body_text = body_text
        # return http_res
            

# 4. Set your class to `py_eureka_client.http_client`. 
http_client.set_http_client(MyHttpClient())

Logger

The default logger is try to write logs to the screen, you can specify the logger handler to write it to a file.

import py_eureka_client.logger as logger
import logging

_formatter = logging.Formatter(fmt='[%(asctime)s]-[%(name)s]-%(levelname)-4s: %(message)s')
_handler = logging.TimedRotatingFileHandler("/var/log/py-eureka-client.log", when="midnight", backupCount=7)
_handler.setFormatter(_formatter)
_handler.setLevel("INFO")

logger.set_handler(_handler)

If you want to add a handler rather than replace the inner one, you can use:

logger.add_handler(_handler)

If you want to change the logger level:

logger.set_level("DEBUG")

This logger will first save all the log record to a global queue, and then output them in a background thread, so it is very suitable for getting several logger with a same handler, especialy the TimedRotatingFileHandler which may slice the log files not quite well in a mutiple thread environment.

Costom Logger

If you want to use your own logger, you can do this:

import py_eureka_client.logger as logger

logger.set_custom_logger(your_logger)

Amazon Data Center Support

This component should support deploying in Amazone EC2, it should automatically load metadata from Amazon metadata service. All the metadata keys come from com.netflix.appinfo.AmazonInfo in Netflix's java client. BUT for the reason that I have no amazon environment to test, it may not work. If errors occurs, please submit an issue and provide some detail logs, I will try to fix it as far as I can. If it works, a reply in this issue is wellcomed.

More Infomation

You can find more information in the project comments.