Merge branch 'main' of https://github.com/MicrosoftDocs/azure-docs-pr into vnet-peer-update

Author: asudbring

articles/cosmos-db/vector-database.md

@@ -117,14 +117,14 @@ DiskANN enables you to perform highly accurate, low latency queriers at any scal
 
 #### Links and samples
 
+- [What is the database behind ChatGPT? - Microsoft Mechanics](https://www.youtube.com/watch?v=6IIUtEFKJec)
 - [Vector indexing in Azure Cosmos DB for NoSQL](index-policy.md#vector-indexes)
 - [VectorDistance system function NoSQL queries](nosql/query/vectordistance.md)
 - [How to setup vector database capabilities in Azure Cosmos DB NoSQL](nosql/vector-search.md)
 - [Python notebook tutorial](https://github.com/microsoft/AzureDataRetrievalAugmentedGenerationSamples)
 - [C# Solution accelerator for building AI apps](https://aka.ms/BuildModernAiAppsSolution)
 - [C# Azure Cosmos DB Chatbot with Azure OpenAI](https://aka.ms/cosmos-chatgpt-sample)
 
-
 ### API for MongoDB
 
 Use the natively [integrated vector database in Azure Cosmos DB for MongoDB](mongodb/vcore/vector-search.md) (vCore architecture), which offers an efficient way to store, index, and search high-dimensional vector data directly alongside other application data. This approach removes the necessity of migrating your data to costlier alternative vector databases and provides a seamless integration of your AI-driven applications.
@@ -150,19 +150,6 @@ Use the natively integrated vector database in [Azure Cosmos DB for PostgreSQL](
 #### Code sample
 - Python: [Python notebook tutorial - food review chatbot](https://github.com/microsoft/AzureDataRetrievalAugmentedGenerationSamples/tree/main/Python/CosmosDB-PostgreSQL_CognitiveSearch)
 
-### NoSQL API
-
-> [!NOTE]
-> For our NoSQL API, the native integration of a state-of-the-art vector indexing algorithm will be announced during Build in May 2024. Please stay tuned.
-
-The natively integrated vector databaseg in the NoSQL API is under development. In the meantime, you may implement RAG patterns with Azure Cosmos DB for NoSQL and [Azure AI Search](../search/vector-search-overview.md). This approach enables powerful integration of your data residing in the NoSQL API into your AI-oriented applications.
-
-#### Links & Code samples
-
-- [What is the database behind ChatGPT? - Microsoft Mechanics](https://www.youtube.com/watch?v=6IIUtEFKJec)
-- [.NET tutorial - Build and Modernize AI Applications](https://github.com/Azure/Build-Modern-AI-Apps-Hackathon)
-- [.NET tutorial - Bring Your Data to ChatGPT](https://github.com/Azure/Vector-Search-AI-Assistant/tree/cognitive-search-vector)
-- [Azure Data + RAG samples with Azure OpenAI](https://github.com/microsoft/AzureDataRetrievalAugmentedGenerationSamples/)
 ### Next step 
 
 [30-day Free Trial without Azure subscription](https://azure.microsoft.com/try/cosmosdb/)

articles/expressroute/expressroute-about-virtual-network-gateways.md

@@ -76,7 +76,7 @@ Before you create an ExpressRoute gateway, you must create a gateway subnet. The
 
 When you create the gateway subnet, you specify the number of IP addresses that the subnet contains. The IP addresses in the gateway subnet are allocated to the gateway VMs and gateway services. Some configurations require more IP addresses than others. 
 
-When you're planning your gateway subnet size, refer to the documentation for the configuration that you're planning to create. For example, the ExpressRoute/VPN Gateway coexist configuration requires a larger gateway subnet than most other configurations. Further more, you might want to make sure your gateway subnet contains enough IP addresses to accommodate possible future configurations. While you can create a gateway subnet as small as /29, we recommend that you create a gateway subnet of /27 or larger (/27, /26 etc.). If you plan on connecting 16 ExpressRoute circuits to your gateway, you **must** create a gateway subnet of /26 or larger. If you're creating a dual stack gateway subnet, we recommend that you also use an IPv6 range of /64 or larger. This set up accommodates most configurations.
+When you're planning your gateway subnet size, refer to the documentation for the configuration that you're planning to create. For example, the ExpressRoute/VPN Gateway coexist configuration requires a larger gateway subnet than most other configurations. Further more, you might want to make sure your gateway subnet contains enough IP addresses to accommodate possible future configurations. We recommend that you create a gateway subnet of /27 or larger (/27, /26 etc.). If you plan on connecting 16 ExpressRoute circuits to your gateway, you **must** create a gateway subnet of /26 or larger. If you're creating a dual stack gateway subnet, we recommend that you also use an IPv6 range of /64 or larger. This set up accommodates most configurations.
 
 The following Resource Manager PowerShell example shows a gateway subnet named GatewaySubnet. You can see the CIDR notation specifies a /27, which allows for enough IP addresses for most configurations that currently exist.
 

articles/search/search-create-service-portal.md

@@ -223,7 +223,7 @@ Azure AI Search restricts the [number of search services](search-limits-quotas-c
 
 You must have Owner or Contributor permissions on the subscription to request quota.
 
-Maximum quota for a given tier and region combination is an extra 100 search services over the baseline quota (which means 106, 108, or 116 [depending on the tier](search-limits-quotas-capacity.md#subscription-limits)). You can't increase quota for the Free tier.
+Maximum quota for a given tier and region combination is an extra 100 search services over the baseline quota (which means 106, 108, or 116 [depending on the tier](search-limits-quotas-capacity.md#subscription-limits)). For more than 100, file a support ticket. You can't increase quota for the Free tier.
 
 1. Sign in to the Azure portal, search for "quotas" in your dashboard, and then select the **Quotas** service.
 

articles/search/search-get-started-portal-image-search.md

@@ -44,7 +44,9 @@ Sample data consists of image files in the [azure-search-sample-data](https://gi
 
 All of the above resources must have public access enabled for the portal nodes to be able to access them. Otherwise, the wizard fails. After the wizard runs, firewalls and private endpoints can be enabled on the different integration components for security.
 
-A free search service supports role-based access control on connections to Azure AI Search, but it doesn't support managed identities on outbound connections to Azure Storage or Azure AI Vision. This means you must use key-based authentication on free search service connections to other Azure services. For more secure connections, use basic tier or above and [configure a managed identity](search-howto-managed-identities-data-sources.md) and role assignments to admit requests from Azure AI Search on other Azure services. 
+If private endpoints are already present and can't be disabled, the alternative option is to run the respective end-to-end flow from a script or program from a virtual machine within the same virtual network as the private endpoint. Here's a [Python code sample](https://github.com/Azure/azure-search-vector-samples/tree/main/demo-python/code/integrated-vectorization) for integrated vectorization. In the same [GitHub repo](https://github.com/Azure/azure-search-vector-samples/tree/main) are samples in other programming languages. 
+
+A free search service supports role-based access control on connections to Azure AI Search, but it doesn't support managed identities on outbound connections to Azure Storage or Azure AI Vision. This means you must use key-based authentication on free search service connections to other Azure services. For more secure connections, use basic tier or higher and [configure a managed identity](search-howto-managed-identities-data-sources.md) and role assignments to admit requests from Azure AI Search on other Azure services. 
 
 ## Check for space
 

articles/search/search-get-started-portal-import-vectors.md

@@ -9,7 +9,7 @@ ms.service: cognitive-search
 ms.custom:
   - build-2024
 ms.topic: quickstart
-ms.date: 05/30/2024
+ms.date: 06/17/2024
 ---
 
 # Quickstart: Import and vectorize data wizard (preview)
@@ -19,10 +19,17 @@ ms.date: 05/30/2024
 
 Get started with [integrated vectorization (preview)](vector-search-integrated-vectorization.md) using the **Import and vectorize data** wizard in the Azure portal. This wizard calls a user-specified embedding model to vectorize content during indexing and for queries.
 
-In this preview version of the wizard:
+You need three Azure resources and some sample files to complete this walkthrough:
 
-+ Source data is either blobs in Azure Storage or files in OneLake, using the default parsing mode (one search document per blob or file).
-+ Index schema is nonconfigurable. Source fields include `content` (chunked and vectorized), `metadata_storage_name` for title, and a `metadata_storage_path` for the document key, represented as `parent_id` in the Index.
+> [!div class="checklist"]
+> + Azure Blob storage or Microsoft Fabric with OneLake for your data
+> + Azure vectorizations: either Azure AI services multiservice account, Azure OpenAI, or Azure AI Studio model catalog
+> + Azure AI Search for indexing and queries
+
+## Preview limitations
+
++ Source data is either Azure Blob Storage or OneLake files and shortcuts, using the default parsing mode (one search document per blob or file).
++ Index schema is nonconfigurable. Source fields include "content" (chunked and vectorized), "metadata_storage_name" for title, and a "metadata_storage_path" for the document key, represented as `parent_id` in the Index.
 + Chunking is nonconfigurable. The effective settings are:
 
   ```json
@@ -31,27 +38,29 @@ In this preview version of the wizard:
   pageOverlapLength: 500
   ```
 
-For more configuration and data source options, try Python or the REST APIs. See [integrated vectorization sample](https://github.com/Azure/azure-search-vector-samples/blob/main/demo-python/code/integrated-vectorization/azure-search-integrated-vectorization-sample.ipynb) for details.
+For fewer limitations or more data source options, try a code-base approach. See [integrated vectorization sample](https://github.com/Azure/azure-search-vector-samples/blob/main/demo-python/code/integrated-vectorization/azure-search-integrated-vectorization-sample.ipynb) for details.
 
 ## Prerequisites
 
 + An Azure subscription. [Create one for free](https://azure.microsoft.com/free/).
 
-+ Azure AI Search, in any region and on any tier, with two caveats:
++ For data, use either an [Azure Storage account](/azure/storage/common/storage-account-overview) or a [OneLake lakehouse](search-how-to-index-onelake-files.md). For Azure Storage, use a standard performance (general-purpose v2) account. Access tiers can be hot, cool, and cold.
+
++ For vectorization, have an Azure AI services multiservice account or [Azure OpenAI](https://aka.ms/oai/access) endpoint with deployments.
 
-  First, role-based access control isn't available on the free tier. Basic tier and higher  provide role-based access control, which is required for *OneLake indexing* and recommended for connections to embedding models.
+  For [multimodal with Azure AI Vision](/azure/ai-services/computer-vision/how-to/image-retrieval), create an Azure AI service in SwedenCentral, EastUS, NorthEurope, WestEurope, WestUS, SoutheastAsia, KoreaCentral, FranceCentral, AustraliaEast, WestUS2, SwitzerlandNorth, JapanEast. [Check the documentation](/azure/ai-services/computer-vision/how-to/image-retrieval?tabs=csharp) for an updated list.
 
-  Second, for multimodal embeddings with Azure AI Vision or image-related transformations, your search service must be in the *same region* as Azure AI Vision. Currently, those regions are: SwedenCentral, EastUS, NorthEurope, WestEurope, WestUS, SoutheastAsia, KoreaCentral, FranceCentral, AustraliaEast, WestUS2, SwitzerlandNorth, JapanEast. [Check the documentation](/azure/ai-services/computer-vision/how-to/image-retrieval?tabs=csharp) for an updated list.
+  You can also use [Azure AI Studio model catalog](/azure/ai-studio/what-is-ai-studio) (and hub and project) with model deployments.
 
-+ A supported embedding model: [Azure OpenAI](https://aka.ms/oai/access) endpoint with deployments, [Azure AI Vision](/azure/ai-services/computer-vision/how-to/image-retrieval) in a supported region, or [Azure AI Studio model catalog](/azure/ai-studio/what-is-ai-studio) (and hub and project) with model deployments.
++ Azure AI Search, in the same region as your Azure AI service. We recommend Basic tier or higher.s
 
-+ A supported data source: [Azure Storage account](/azure/storage/common/storage-account-overview) or a [OneLake lakehouse](search-how-to-index-onelake-files.md). For Azure Storage, use a standard performance (general-purpose v2) account. Access tiers can be hot, cool, and cold.
++ Role assignments or API keys are required for connections to embedding models and data sources. Instructions for role-based access are provided in this article.
 
-+ Role assignments or API keys are required for connections to embedding models and data sources. Instructions are provided in this article.
+All of the above resources must have public access enabled for the portal nodes to be able to access them. Otherwise, the wizard fails. After the wizard runs, firewalls and private endpoints can be enabled on the different integration components for security.
 
-+ All components (data source and embedding endpoint) must have public access enabled for the portal nodes to be able to access them. Otherwise, the wizard fails. After the wizard runs, firewalls and private endpoints can be enabled on the different integration components for security. 
+If private endpoints are already present and can't be disabled, the alternative option is to run the respective end-to-end flow from a script or program from a virtual machine within the same virtual network as the private endpoint. Here's a [Python code sample](https://github.com/Azure/azure-search-vector-samples/tree/main/demo-python/code/integrated-vectorization) for integrated vectorization. In the same [GitHub repo](https://github.com/Azure/azure-search-vector-samples/tree/main) are samples in other programming languages. 
 
-   If private endpoints are already present and can't be disabled, the alternative option is to run the respective end-to-end flow from a script or program from a virtual machine within the same virtual network as the private endpoint. Here's a [Python code sample](https://github.com/Azure/azure-search-vector-samples/tree/main/demo-python/code/integrated-vectorization) for integrated vectorization. In the same [GitHub repo](https://github.com/Azure/azure-search-vector-samples/tree/main) are samples in other programming languages. 
+A free search service supports role-based access control on connections to Azure AI Search, but it doesn't support managed identities on outbound connections to Azure Storage or Azure AI Vision. This means you must use key-based authentication on free search service connections to other Azure services. For more secure connections, use basic tier or above and [configure a managed identity](search-howto-managed-identities-data-sources.md) and role assignments to admit requests from Azure AI Search on other Azure services.
 
 ## Check for space
 

articles/search/vector-search-ranking.md

@@ -32,7 +32,7 @@ Only vector fields marked as `searchable` in the index, or as `searchFields` in
 
 Exhaustive KNN calculates the distances between all pairs of data points and finds the exact `k` nearest neighbors for a query point. It's intended for scenarios where high recall is of utmost importance, and users are willing to accept the trade-offs in query latency. Because it's computationally intensive, use exhaustive KNN for small to medium datasets, or when precision requirements outweigh query performance considerations. 
 
-A seconary use case is to build a dataset to evaluate approximate nearest neighbor algorithm recall. Exhaustive KNN can be used to build the ground truth set of nearest neighbors.
+A secondary use case is to build a dataset to evaluate approximate nearest neighbor algorithm recall. Exhaustive KNN can be used to build the ground truth set of nearest neighbors.
 
 Exhaustive KNN support is available through [2023-11-01 REST API](/rest/api/searchservice/search-service-api-versions#2023-11-01), [2023-10-01-Preview REST API](/rest/api/searchservice/search-service-api-versions#2023-10-01-Preview), and in Azure SDK client libraries that target either REST API version.