Elasticsearch is an open source, distributed, search and analytics RESTful engine. Spring Boot provides basic autoconfiguration for Elasticsearch clients.
Spring Boot supports several clients:
-
Official low-level Java clients and high-level REST clients
-
ReactiveElasticsearchClient
provided by Spring Data Elasticsearch
Spring Boot provides a special "starter" called spring-boot-starter-data-elasticsearch
.
Connecting to Elasticsearch via REST clients
Elasticsearch comes with two different REST clients that can be used to send requests to the
cluster: a low-level client from the org.elasticsearch.client:elasticsearch-rest-client
module and a
high-level client from the org.elasticsearch.client:elasticsearch-high-level-client
module. In
addition, Spring Boot provides support for a reactive client based on the WebClient
from the Spring
Framework, from the org.springframework.data:spring-data-elasticsearch
module. By default, clients will
target localhost:9200
. You can use the spring.elasticsearch.*
properties to further fine-tune the client configuration, as shown in the following example:
spring.elasticsearch.uris=https://search.example.com:9200
spring.elasticsearch.socket-timeout=10s
spring.elasticsearch.username=user
spring.elasticsearch.password=secret
spring:
elasticsearch:
uris: "https://search.example.com:9200"
socket-timeout: "10s"
username: "user"
password: "secret"
Connecting to Elasticsearch via RestClient
If elasticsearch-rest-client
is in the classpath, Spring Boot will automatically configure and register
the RestClient
bean. If elasticsearch-rest-high-level-client
is in the classpath, the
RestHighLevelClient
bean will also be automatically configured. Due to the deprecation of RestHighLevelClient
by Elasticsearch, its autoconfiguration has been deprecated and will be removed in a future release. In addition to
the properties described earlier, you can register an arbitrary number of beans that implement RestClientBuilderCustomizer
for more detailed customization of RestClient
and RestHighLevelClient
. To gain full
control over client configuration, define the RestClientBuilder
bean.
Additionally, if elasticsearch-rest-client-sniffer
is in the classpath, Sniffer
will
automatically be configured to automatically discover nodes from a running Elasticsearch cluster and install them
into a bin RestClient
. You can further customize the Sniffer
configuration, as shown in
the following example:
spring.elasticsearch.restclient.sniffer.interval=10m
spring.elasticsearch.restclient.sniffer.delay-after-failure=30s
spring:
elasticsearch:
restclient:
sniffer:
interval: "10m"
delay-after-failure: "30s"
Connecting to Elasticsearch via ReactiveElasticsearchClient
Spring Data Elasticsearch contains a ReactiveElasticsearchClient
for querying Elasticsearch instances in
a reactive manner. It is a plugin for the WebClient
from WebFlux, so the dependencies spring-boot-starter-elasticsearch
and spring-boot-starter-webflux
are useful to enable support tools code>.
By default, Spring Boot will automatically configure and register the ReactiveElasticsearchClient
. In
addition to the properties described previously, the spring.elasticsearch.webclient.*
properties can be
used to configure reactive programming-specific options, as shown in the following example:
spring.elasticsearch.webclient.max-in-memory-size=1MB
spring:
elasticsearch:
webclient:
max-in-memory-size: "1MB"
If the configuration properties of spring.elasticsearch.
and spring.elasticsearch.webclient.
are not enough and you want fully manage the client configuration, you can register a custom bean ClientConfiguration
.
Connecting to Elasticsearch via Spring Data
To connect to Elasticsearch, a RestHighLevelClient
bean must be defined, which is automatically
configured by Spring Boot or passed manually by the application (see previous sections). With this configuration,
ElasticsearchRestTemplate
can be implemented just like any other Spring bean, as shown in the following
example:
@Component
public class MyBean {
private final ElasticsearchRestTemplate template;
public MyBean(ElasticsearchRestTemplate template) {
this.template = template;
}
}
@Component
class MyBean(private val template: ElasticsearchRestTemplate) {
}
Given spring-data-elasticsearch
and the necessary dependencies to use WebClient
(usually
spring-boot-starter-webflux
), Spring Boot can also automatically configure ReactiveElasticsearchClient
and ReactiveElasticsearchTemplate
as beans. They are the reactive equivalent of other REST clients.
Spring Data Elasticsearch Repositories
Spring Data includes repository support for Elasticsearch. As with the JPA repositories discussed earlier, the basic principle is that queries are built automatically based on method names.
In fact, both Spring Data JPA and Spring Data Elasticsearch share the same general infrastructure. You can take the
JPA example from the previous article and, provided that City
is now an Elasticsearch class with the
@Document
annotation, rather than a JPA object with the @Entity
annotation, it will work
exactly the same.
Spring Boot supports both classic and reactive Elasticsearch repositories through the
ElasticsearchRestTemplate
or ReactiveElasticsearchTemplate
beans. Most likely, these beans
will be automatically configured by Spring Boot if the necessary dependencies are present.
If you want to use your own template to provide support for Elasticsearch repositories, you can add your own ElasticsearchRestTemplate
or ElasticsearchOperations
via a bean marked with the @Bean
if its name is "elasticsearchTemplate"
.
The same applies to ReactiveElasticsearchTemplate
and ReactiveElasticsearchOperations
, but
the bean name must be "reactiveElasticsearchTemplate"
.
You can disable repository support using the following property:
spring.data.elasticsearch.repositories.enabled=false
spring:
data:
elasticsearch:
repositories:
enabled: false
GO TO FULL VERSION