如何在 supervisord 中捕获 Flask CLI 命令的输出并将其重定向到标准输出？答案

【问题标题】：How to capture output of Flask CLI command in supervisord and redirect it to stdout?如何在 supervisord 中捕获 Flask CLI 命令的输出并将其重定向到标准输出？
【发布时间】：2020-06-16 15:08:32
【问题描述】：

这是我的 Flask CLI 脚本。它等待来自rabbitmq 的消息并打印出来。

import pika
import click
from app import app_queue
from app import RABBITMQ_USER
from app import RABBITMQ_HOST
from app import RABBITMQ_PASSWORD
from app import RABBITMQ_PASSWORD
from app import RABBITMQ_PORT
from app import RABBITMQ_VHOST
from flask import current_app
from flask.cli import with_appcontext
from threading import Event

credentials = pika.PlainCredentials(RABBITMQ_USER, RABBITMQ_PASSWORD)
parameters = pika.ConnectionParameters(RABBITMQ_HOST, RABBITMQ_PORT, RABBITMQ_VHOST, credentials)
connection = pika.BlockingConnection(parameters)
channel = connection.channel()


@app_queue.command("listen", help="Listen to default queue.")
@click.argument("queue", required=False)
@with_appcontext
def consume_message(queue):
    queue = queue if queue else 'results'
    print(f'Command is listening for "{queue}" queue.', flush=True)

    channel.queue_declare(queue, durable=True, exclusive=False)
    channel.basic_consume(queue=queue, on_message_callback=process_results)

    print(f'[*] Waiting for messages. from "{queue}" queue. To exit press CTRL+C', flush=True)
    channel.start_consuming()

    Event().wait()


def process_results(ch, method, properties, body):
    print(" [x] Received %r" % body, flush=True)
    ch.basic_ack(delivery_tag = method.delivery_tag)

脚本由supervisord 执行，触发flask queue listen 命令。

[program:app_rabbitmq]
user=root
numprocs=1
directory=/src
autostart=true
autorestart=true
process_name=%(program_name)s
redirect_stderr=true
stdout_logfile=/dev/stdout
stdout_logfile_maxbytes=0
command=flask queue listen

如何将 Flask CLI 命令的输出重定向到标准输出，以便使用docker logs <my_container> 命令查看？

【问题讨论】：

标签： python docker flask supervisord

【解决方案1】：

我找到了解决方案。这是我做的三件事。

我将日志重定向到/dev/stdout，而是将其重定向到/proc/1/fd/1
我以 nodaemon 模式启动 supervisord supervisord -n
我在我的docker镜像export PYTHONUNBUFFERED=True和export PYTHONIOENCODING=UTF-8中设置了这两个环境变量

【讨论】：