请求超时/延迟问题

Question

我的应用程序使用 Django REST 框架发布到 Instagram。发布到 Instagram 是一个两步过程。首先，您必须将内容发送到媒体容器。然后你必须等到 Instagram 处理完照片/视频。处理完媒体后，您将创建 ID 发送到 Instagram 以公开发布帖子。为了保证 Instagram 在发送创建 ID 之前有足够的时间处理媒体，我使用 Django 的 time.sleep 函数将请求暂停 60 秒。问题是这适用于我的桌面，但在我的 ec2 实例上，当我暂停请求超过 10 秒时，相同的代码会失败。为了运行我的应用程序，我使用了 amazon EC2、Gunicorn 和 nginx。

我的 Django 代码是这样的

```#print('posting photo to facebook')
    video_url = data.get('video_url')
    container_url = "https://graph.facebook.com/" + page_id + "/media?media_type=VIDEO&\
    video_url=" + video_url + "&caption=" + text + "&access_token=" + access_token
    res = requests.post(container_url)
    data = res.json()
    time.sleep(60)
    publish_url = "https://graph.facebook.com/" + page_id + "/media_publish?\
    creation_id=" + data['id'] + "&access_token=" + access_token
    res = requests.post(publish_url)
    data = res.json()
    print('result of video posting attempt')
    print(json.dumps(data));
    a = res.status_code
    b = 200
    #print(res.text)
    if a != b:
        return Response(data = {"data": {"id":"" , "post_id":"" }})
    return Response(data = data, status = res.status_code)```

我的 gunicorn.service 文件

       Description=gunicorn daemon
       Requires=gunicorn.socket
       After=network.target
      
       [Service]
       User=ubuntu
       Group=ubuntu
       WorkingDirectory=/home/ubuntu/xxxxx-xxxxxx-xxxxxx
       ExecStart=/home/ubuntu/xxxxx-xxxxxx-xxxxxx/xxxxxxx/bin/gunicorn \
       --access-logfile - \
       --workers 3 \
       --bind unix:/run/gunicorn.sock \
       xxxxxxxxxx.wsgi:application
       TimeoutStopSec=120

      [Install]
      WantedBy=multi-user.target

我的 nginx 配置文件

       listen 80;
       listen [::]:80;
       server_name xxxxxxxxx.com *.xxxxxxxxx.com;
       return 301 https://$host$request_uri;
       }

      server {
      listen 443 ssl http2;
      listen [::]:443 ssl http2;
      server_name xxxxxxxxx.com *.xxxxxxxxx.com;

      #SSL/TLS settings
      ssl_certificate /etc/letsencrypt/live/xxxxxxxxx.com/fullchain.pem;
      ssl_certificate_key /etc/letsencrypt/live/xxxxxxxxx.com/privkey.pem;
      client_max_body_size 100M;
      proxy_read_timeout 300;
      proxy_connect_timeout 300;
      proxy_send_timeout 300;

      location = /favicon.ico {
        access_log off;
        log_not_found off;
      }

      location /static/ {
        root /home/ubuntu/xxxxxxxxx-xxxxxxxxx-xxxxxxxxx;
      }

      location / {
        include proxy_params;
        proxy_pass http://unix:/run/gunicorn.sock;
      }

      }```

Answer 1

我建议使用celery 处理延迟任务。例如：

@shared_task
def task_make_public(publish_url):
    res = requests.post(publish_url)
    ...

res = requests.post(container_url)
data = res.json()
publish_url = ...

task_make_public.apply_async((publish_url,), eta=now + timedelta(seconds=60))

更简单的设置适用于django-background-tasks。它不像 celery 那样需要 celeworker、celerybeat 和 reddis 来运行任务。

从 PyPI 安装

pip install django-background-tasks

添加到 INSTALLED_APPS：

INSTALLED_APPS = (
    # ...
    'background_task',
    # ...
)

迁移数据库

python manage.py migrate

代码示例

from background_task import background
from django.contrib.auth.models import User

@background
def task_make_public(publish_url):
    res = requests.post(publish_url)
    ...

res = requests.post(container_url)
data = res.json()
publish_url = ...

task_make_public(publish_url, schedule=timedelta(seconds=60)) # 60 seconds from now