コミュニティ メンバーがこの投稿を同様の質問に関連付けています:
Custom Translator:以前動作していたモデルを含めデプロイが失敗しAPIがエラー400002を返す(japaneast)

モデレーターのみがこのコンテンツを編集できます。

Custom Translator: Deployment fails and API returns 400002 even for previously working models (Japan East)

azure translator user 0 評価のポイント
2026-04-06T09:24:12.2133333+00:00

Hello,

I am facing a regression issue with Azure Custom Translator in the Japan East region. As of today (Monday, April 6), both my previously working models and newly trained models are failing to deploy...

Background:

  • Last Friday (April 3): A model I had created earlier was working perfectly. I also trained a new model on the same day but did not test its deployment then...

Today (April 6): Both the "previously working model" and the "newly trained model" fail to deploy...

Symptoms:

  1. Deployment Flakiness: Clicking "Deploy" in the portal shows temporary success, but it changes to "Deployment Failed" after a few minutes. No detailed error is shown in the portal.

Missing Logs: There are no failure records in the Azure Activity Log or Log Analytics (AzureDiagnostics) for these deployment attempts.

API Error (400002): When calling the API with the Category ID, I receive:

Code: 400002

  __Message:__ "The category parameter is invalid"

  
     Standard translation (without the category parameter) works fine with the same key.
```__Environment:__

__Resource:__ Translator (S1 Tier)

__Region:__ Japan East

__Subscription:__ Active (No billing alerts in Cost Management)

Since even a previously verified model is now failing, I suspect a service-side issue or a regional capacity constraint in Japan East that started over the weekend. Has anyone else experienced this, or is there a way to see the "hidden" reason for these deployment failures?

Thank you.

Foundry Tools の Azure Translator
Foundry Tools の Azure Translator

簡単な REST API 呼び出しを使用して簡単に機械翻訳を行う Azure サービス。