HSEARCH-3541 Do not rely on the Elasticsearch client's "max retry tim…

…eout" anymore

It is no longer available in Elasticsearch 7, and we already have our
own timeout. It just won't retry anymore.
See elastic/elasticsearch#25951 (comment)

Note this clears the path for more useful timeouts that start when the
request is sent to the cluster rather than when the request is queued.
See https://hibernate.atlassian.net/browse/HSEARCH-2836.

backend/elasticsearch/src/main/java/org/hibernate/search/backend/elasticsearch/client/impl/ElasticsearchClientFactoryImpl.java

@@ -139,13 +139,6 @@ private RestClient createClient(ConfigurationPropertySource propertySource,
 		ServerUris hosts = ServerUris.fromStrings( HOST.get( propertySource ) );
 
 		return RestClient.builder( hosts.asHostsArray() )
-				/*
-				 * Note: this timeout is currently only used on retries,
-				 * but should we start using the synchronous methods of RestClient,
-				 * it would be applied to synchronous requests too.
-				 * See https://github.com/elastic/elasticsearch/issues/21789#issuecomment-287399115
-				 */
-				.setMaxRetryTimeoutMillis( maxRetryTimeoutMillis )
 				.setRequestConfigCallback( b -> customizeRequestConfig( b, propertySource ) )
 				.setHttpClientConfigCallback(
 						b -> customizeHttpClientConfig( b, httpClientConfigurers, propertySource, hosts )