Azure AI Foundry添加Agent知识库时搜索索引加载失败（400错误）

Unable to load search service indexes Request failed with status code 400.

当前环境状态 #

• Azure AI Search已与项目关联
• 已配置所需权限
• 可在搜索服务中执行搜索操作

索引创建方式 #

使用Terraform通过REST API创建了搜索索引、技能集和索引器，相关代码如下：

# Create Azure AI Search Index
resource "null_resource" "create_search_index" {
  triggers = {
    search_service = azurerm_search_service.example.id
  }

  provisioner "local-exec" {
    command = <<EOT
      # Create index definition JSON
      cat > index.json <<EOF
{
  "name": "cities-index",
  "fields": [
    {
      "name": "chunk_id",
      "type": "Edm.String",
      "key": true,
      "analyzer": "keyword",
      "searchable": true,
      "sortable": true,
      "filterable": false,
      "facetable": false
    },
    {
      "name": "parent_id",
      "type": "Edm.String",
      "filterable": true,
      "searchable": false,
      "sortable": false,
      "facetable": false
    },
    {
      "name": "chunk",
      "type": "Edm.String",
      "searchable": true,
      "filterable": false,
      "sortable": false,
      "facetable": false
    },
    {
      "name": "title",
      "type": "Edm.String",
      "searchable": true,
      "filterable": false,
      "sortable": false,
      "facetable": false
    },
    {
      "name": "text_vector",
      "type": "Collection(Edm.Single)",
      "searchable": true,
      "retrievable": true,
      "filterable": false,
      "sortable": false,
      "facetable": false,
      "stored": true,
      "vectorSearchProfile": "cities-index-text-profile",
      "dimensions": 1536
    },
    {
      "name": "metadata_storage_path",
      "type": "Edm.String",
      "searchable": true,
      "filterable": false,
      "retrievable": false,
      "stored": true,
      "sortable": false,
      "facetable": false,
      "key": false
    }
  ],
  "vectorSearch": {
    "profiles": [
      {
        "name": "cities-index-text-profile",
        "algorithm": "cities-index-algorithm",
        "vectorizer": "cities-index-text-vectorizer"
      }
    ],
    "algorithms": [
      {
        "kind": "hnsw",
        "name": "cities-index-algorithm"
      }
    ],
    "vectorizers": [
      {
        "kind": "azureOpenAI",
        "azureOpenAIParameters": {
          "resourceUri": "https://${azurerm_ai_services.example.name}.openai.azure.com/",
          "deploymentId": "text-embedding-ada-002",
          "modelName": "text-embedding-ada-002",
          "apiKey": "${azurerm_ai_services.example.primary_access_key}"
        },
        "name": "cities-index-text-vectorizer"
      }
    ],
    "compressions": []
  },
  "semantic": {
    "defaultConfiguration": "cities-index-semantic-configuration",
    "configurations": [
      {
        "name": "cities-index-semantic-configuration",
        "prioritizedFields": {
          "titleField": {
            "fieldName": "title"
          },
          "prioritizedContentFields": [
            {
              "fieldName": "chunk"
            }
          ]
        }
      }
    ]
  }
}
EOF

      # Create index using POST method
      curl -X POST \
        -H "Content-Type: application/json" \
        -H "api-key: ${azurerm_search_service.example.primary_key}" \
        -d @index.json \
        "https://${azurerm_search_service.example.name}.search.windows.net/indexes?api-version=2025-05-01-preview"
    EOT

    interpreter = ["bash", "-c"]
  }

  depends_on = [
    azurerm_search_service.example
  ]
}

# Create Azure AI Search Skillset
resource "null_resource" "create_search_skillset" {
  triggers = {
    search_service = azurerm_search_service.example.id
  }

  provisioner "local-exec" {
    command = <<EOT
      # Create skillset definition JSON
      cat > skillset.json <<EOF
{
  "name": "cities-skillset",
  "description": "Skillset to chunk documents and generate embeddings",
  "skills": [
    {
      "@odata.type": "#Microsoft.Skills.Text.SplitSkill",
      "description": "Split skill to chunk documents",
      "textSplitMode": "pages",
      "maximumPageLength": 2000,
      "pageOverlapLength": 500,
      "context": "/document",
      "inputs": [
        {
          "name": "text",
          "source": "/document/content"
        }
      ],
      "outputs": [
        {
          "name": "textItems",
          "targetName": "pages"
        }
      ]
    },
    {
      "@odata.type": "#Microsoft.Skills.Text.AzureOpenAIEmbeddingSkill",
      "context": "/document/pages/*",
      "inputs": [
        {
          "name": "text",
          "source": "/document/pages/*"
        }
      ],
      "outputs": [
        {
          "name": "embedding",
          "targetName": "text_vector"
        }
      ],
      "resourceUri": "https://${azurerm_ai_services.example.name}.openai.azure.com/",
      "deploymentId": "text-embedding-ada-002",
      "modelName": "text-embedding-ada-002",
      "apiKey": "${azurerm_ai_services.example.primary_access_key}",
      "dimensions": 1536
    }
  ],
  "indexProjections": {
    "selectors": [
      {
        "targetIndexName": "cities-index",
        "parentKeyFieldName": "parent_id",
        "sourceContext": "/document/pages/*",
        "mappings": [
          {
            "name": "title",
            "source": "/document/title"
          },
          {
            "name": "chunk",
            "source": "/document/pages/*"
          },
          {
            "name": "text_vector",
            "source": "/document/pages/*/text_vector"
          },
          {
            "name": "metadata_storage_path",
            "source": "/document/metadata_storage_path",
            "inputs": []
          }
        ]
      }
    ],
    "parameters": {
      "projectionMode": "skipIndexingParentDocuments"
    }
  }
}
EOF

      # Create skillset using POST method
      curl -X POST \
        -H "Content-Type: application/json" \
        -H "api-key: ${azurerm_search_service.example.primary_key}" \
        -d @skillset.json \
        "https://${azurerm_search_service.example.name}.search.windows.net/skillsets?api-version=2025-05-01-preview"
    EOT

    interpreter = ["bash", "-c"]
  }

  depends_on = [
    azurerm_search_service.example,
    azurerm_ai_services.example
  ]
}

# Create Azure AI Search Indexer
resource "null_resource" "create_search_indexer" {
  triggers = {
    search_service = azurerm_search_service.example.id
  }

  provisioner "local-exec" {
    command = <<EOT
      # Create indexer definition JSON
      cat > indexer.json <<EOF
{
  "name": "cities-indexer",
  "targetIndexName": "cities-index",
  "dataSourceName": "azureblob-1752751172828-datasource",
  "skillsetName": "cities-skillset",
  "schedule": null,
  "fieldMappings": [
    {
      "sourceFieldName": "metadata_storage_name",
      "targetFieldName": "title"
    }
  ],
  "parameters": {
    "configuration": {
      "dataToExtract": "contentAndMetadata",
      "parsingMode": "default"
    }
  }
}
EOF

      # Create indexer using POST method
      curl -X POST \
        -H "Content-Type: application/json" \
        -H "api-key: ${azurerm_search_service.example.primary_key}" \
        -d @indexer.json \
        "https://${azurerm_search_service.example.name}.search.windows.net/indexers?api-version=2025-05-01-preview"
    EOT

    interpreter = ["bash", "-c"]
  }

  depends_on = [
    null_resource.create_search_datasource,
    null_resource.create_search_index,
    null_resource.create_search_skillset
  ]
}

内容的提问来源于stack exchange，提问作者One Developer