如何监控Bucket中文件提取作业进度？除轮询API外是否有其他方案？

Great question! Let me start by clearing up a misunderstanding you had about the Webhooks API first:

Your initial assumption that Webhooks only work with files in folders is incorrect. The Webhooks API does support tracking extraction jobs for files stored directly in Buckets.

You can set up Webhooks to listen for extraction-related events (like extraction.finished or extraction.failed) by configuring targeted filters. When creating a Webhook, you can specify your Bucket's URN or individual file URNs as the resource filter, and set the event filter to extraction.* to capture all extraction-related updates. This way, you’ll get real-time notifications without having to manually poll the API.

Now, besides polling the GET :urn/manifest endpoint, here are two solid alternative methods to monitor extraction progress:

• Use the callbackUrl parameter when creating your extraction job
  When you send a POST job request to the Model Derivative API, you can include a callbackUrl in the request body. As soon as the extraction job finishes (whether successfully or with errors), Autodesk’s servers will send a POST notification to this URL with the job status details. This is perfect if you want job-specific alerts without setting up a persistent Webhook.
  Here’s a quick example of what that request body might look like:

  {
    "input": {
      "urn": "your-file-urn-in-bucket"
    },
    "output": {
      "formats": [
        {
          "type": "svf",
          "views": ["2d", "3d"]
        }
      ]
    },
    "callbackUrl": "https://your-server-endpoint.com/handle-extraction-status"
  }
  

• Configure Webhooks with precise filters
  As I mentioned earlier, Webhooks work seamlessly with Bucket-stored files. You can create a Webhook that only triggers for extraction events tied to your specific Bucket. For instance, set the event filter to extraction.finished and extraction.failed to capture success/failure updates, and use the resource filter to target your Bucket’s unique identifier. This approach is ideal if you’re managing multiple extraction jobs and want a centralized way to receive status updates.

Both the callbackUrl and Webhook methods are far more efficient than polling, especially for high-volume workflows. Polling should really only be a fallback if these real-time notification methods aren’t feasible for your setup.

内容的提问来源于stack exchange，提问作者Ivan Aguilar