Azure Nodes

Implemented Azure node suite:

• azure_openai_chat_model
• embeddings_azure_openai
• azure_storage
• azure_cosmos_db
• azure_monitor_http
• azure_ai_search_vector_store

Each node uses secretRef.secretId to resolve credentials server-side.

Credential setup in UI

1. Open node config
2. In secret dropdown click Create credential
3. Choose provider preset (Azure OpenAI, Azure Storage, Azure Cosmos DB, Azure Monitor, Azure AI Search)
4. Paste credential value (raw string or JSON blob)
5. Save and select created secret
6. Click Test Connection

Supported credential value formats

Azure OpenAI

Raw API key:

<azure-openai-api-key>

JSON:

{
  "apiKey": "<azure-openai-api-key>"
}

Bearer token also accepted:

{
  "accessToken": "eyJ..."
}

Azure Storage

Shared key JSON (recommended):

{
  "accountName": "mystorageacct",
  "accountKey": "<base64-account-key>",
  "endpoint": "https://mystorageacct.blob.core.windows.net"
}

SAS token:

sv=...&ss=...&srt=...&sp=...&se=...&sig=...

Bearer token:

Bearer eyJ...

Azure Cosmos DB

Master key JSON:

{
  "masterKey": "<cosmos-base64-key>"
}

or bearer token JSON:

{
  "accessToken": "eyJ..."
}

Azure Monitor

Bearer token JSON:

{
  "accessToken": "eyJ..."
}

Azure AI Search

Raw API key:

<search-api-key>

JSON:

{
  "apiKey": "<search-api-key>"
}

Node-level operations

Azure Storage

• list_containers
• list_blobs
• get_blob_text
• put_blob_text
• delete_blob

Azure Cosmos DB

• query_items
• read_item
• create_item
• upsert_item
• delete_item

Azure Monitor

• query_logs
• query_metrics
• custom_request

Azure AI Search Vector Store

• vector_search
• upsert_documents
• delete_documents

Demo fallback

Most Azure connectors support useDemoFallback=true. If credential/config is missing, the node can return a deterministic demo response instead of hard failing.