Walkthrough: Configure Azure Cognitive Search

Current version: 9.1
Caution

Azure Cognitive Search will be discontinued in the future and Sitecore will no longer provide support for this service in future releases.

To use Azure Search with Sitecore you must first configure your Azure Search service with your Sitecore instance. 

Format the connection string

The default connection string name for Cloud Search is cloud.search and it contains the following information:

  • serviceUrl – the HTTPS URL of the search service API (for example, https://dk-test.search.windows.net).

  • apiVersion – follows a date format, for example, 2017-11-11 (see more information on API versions).

  • apiKey – the admin key to the service that you obtain from the Azure management portal.

Note

For Sitecore version 9.1 and later, set the API version to 2017-11-11. For information about other supported versions, refer to the Azure Search compatibility table.

The connection string format is:

RequestResponse
<add name="cloud.search" connectionString="serviceUrl=<url>;apiVersion=<apiVersion>;apiKey=<apiKey>" /> 
Note

Azure Cognitive Search for xConnect does not support this feature.

Geo-replicated scenarios

Sitecore supports a Search service with geo-replicated scenarios. To use this type of scenario:

  1. Create two or more Search service instances.

  2. Add connection strings with a pipe separator (|). If you have two search services, for example, searchservice1 and searchservice2, and you want to use them in a geo-replicated scenario, you must use the following connection string:

    RequestResponse
    <add name="cloud.search" 
    connectionString="serviceUrl=https://searchservice1.search.windows.net;apiVersion=2015-02-28;apiKey=AdminKey1|serviceUrl=https://searchservice2.search.windows.net;apiVersion=2015-02-28;apiKey=AdminKey2" /> 

Advanced scaling scenarios

For advanced scaling scenarios, it is best practice to use a dedicated search service for your index. To set this up:

  1. Add a new connection string (for example, cloud.search.analytics).

  2. Configure the corresponding index with this connection string name:

Connection string name

Specify Azure as your search provider

By default Sitecore is distributed with the Solr search provider enabled. 

In order to use Azure instead, you must specify Azure as your search provider in the web.config file.

Rebuild the indexes

You must rebuild the indexes to ensure Sitecore is fully operational.

To rebuild the indexes:

  1. Go to the Sitecore login page (http:// {your_instance}/sitecore/login) and log in with your admin credentials.

  2. On the Sitecore Launchpad, click Control Panel, and select Indexing manager.

  3. To select and rebuild all of the indexes, on the Indexing Manager page, click Select all, Rebuild.

    Note

    Rebuilding the index is a time consuming operation and can take 15 minutes or more.

    When the indexes are rebuilt, the Sitecore search indexes appear in the Search service window of the Azure Portal.

    Indexing Manager page

Map the Azure field types

Azure Search uses the following Entity Data Model (EDM) field types. Refer to the table to ensure you map the field types correctly.

Field type

Description

Edm.String

Optional: Text that can be tokenized for full-text search, for example, word-breaking, stemming, and so on.

Collection(Edm.String)

A list of strings that can be tokenized for full-text search. There is no upper limit on the number of items in a collection, but remember that the 16 MB upper limit on payload size applies to collections.

Edm.Boolean

Contains true or false values.

Edm.Int32

Contains 32-bit integer values.

Edm.Int64

Contains 64-bit integer values.

Edm.Double

Uses double-precision numeric data.

Edm.DateTimeOffset

The date and time values are represented in the OData V4 format, for example:

yyyy-MM-ddTHH:mm:ss.fffZ

yyyy-MM-ddTHH:mm:ss.fff[+|-]HH:mm.

Note

The precision of DateTime fields is limited to milliseconds. If you upload DateTime values that have submillisecond precision, the returned value will be rounded up to milliseconds, for example:

2015-04-15T10:30:09.7552052Z will be returned as:

2015-04-15T10:30:09.7550000Z.

To see how to define the different types of mapping between .Net and Sitecore, go to the sitecore\contentsearch\indexConfigurations\defaultCloudIndexConfiguration\CloudTypeMapper node in the App_Config\Sitecore\ContentSearch.Azure\Sitecore.ContentSearch.Azure.DefaultIndexConfiguration.config file.

Note

You must define the map element for all custom fields like this:

RequestResponse
<map type="<Field type>" cloudType="<Edm type from list of supported types>" /> 

Map fields

When listing all of the fields that the components use under the index configuration section, it is best practice that you list them as follows:

  • sitecore\contentSearch\configuration\indexes\index\configuration\fieldMap

The supported attributes are:

Attribute

Description

fieldName

The name of the field, as defined for Lucene and Solr.

cloudFieldName

The field name as defined for Cloud. It can only contain letters, numbers, and underscores.

Note

The first character must be a letter.

boost

Use the boost attribute to give one field more importance than another.

format

You must configure this attribute for DateTime type fields. The supported value is yyyy-MM-ddTHH:mm:ss.fffZ.

settingType

The settingType property must be: Sitecore.ContentSearch.Azure.CloudSearchFieldConfiguration, Sitecore.ContentSearch.Azure

Use the support reference for Azure Search to ensure that you configure query support and facet support correctly.

Configure a number of open HTTP connections

You can manage the number of connections that the Azure Search provider uses with the following setting:

RequestResponse
<setting name="ContentSearch.Azure.ServicePoint.ConnectionLimit" value="100" />

In the following scenarios you must adjust the setting value:

  • Decrease the value - If there are other applications that are using HTTP intensively.

  • Increase the value - If an application is under a high load, it takes too long to create new connections or creation fails. This is due to the exceed limit being set too low.

Do you have some feedback for us?

If you have suggestions for improving this article,