Spring configurable, high performance, metered http client instances - spring

Coming from DropWizard I am used to its HttpClientConfiguration and I am baffled that in Spring Boot I cannot find some support for controlling in a similar manner http clients instances to be used, by RestTemplates for example.
To work in production the underlying client implementation should be high performance (e.g. non blocking io, with connection reuse and pooling).
Then I need to set timeouts or authentication, possibly metrics gathering, cookie settings, SSL certificates settings.
All of the above should be easy to set up in different variants for different instances to be used in different contexts (e.g. for service X use these settings and this pool, for Y use another pool and settings) and most parameters should be set via environment-specific properties to have different values in production/qa/development.
Is there something that can be used towards this end?

Below is an example of configuring a HttpClient with a configuration class. It configures basic authentication for all requests through this RestTemplate as well as some tweaks to the pool.
HttpClientConfiguration.java
#Configuration
public class HttpClientConfiguration {
private static final Logger log = LoggerFactory.getLogger(HttpClientConfiguration.class);
#Autowired
private Environment environment;
#Bean
public ClientHttpRequestFactory httpRequestFactory() {
return new HttpComponentsClientHttpRequestFactory(httpClient());
}
#Bean
public RestTemplate restTemplate() {
RestTemplate restTemplate = new RestTemplate(httpRequestFactory());
restTemplate.setInterceptors(ImmutableList.of((request, body, execution) -> {
byte[] token = Base64.encodeBase64((format("%s:%s", environment.getProperty("fake.username"), environment.getProperty("fake.password"))).getBytes());
request.getHeaders().add("Authorization", format("Basic %s", new String(token)));
return execution.execute(request, body);
}));
return restTemplate;
}
#Bean
public HttpClient httpClient() {
PoolingHttpClientConnectionManager connectionManager = new PoolingHttpClientConnectionManager();
// Get the poolMaxTotal value from our application[-?].yml or default to 10 if not explicitly set
connectionManager.setMaxTotal(environment.getProperty("poolMaxTotal", Integer.class, 10));
return HttpClientBuilder
.create()
.setConnectionManager(connectionManager)
.build();
}
/**
* Just for demonstration
*/
#PostConstruct
public void debug() {
log.info("Pool max total: {}", environment.getProperty("poolMaxTotal", Integer.class));
}
}
and an example application.yml
fake.username: test
fake.password: test
poolMaxTotal: 10
You can externalise configuration values to your application.yml as with poolMaxTotal etc. above.
To support different values per environment you can use Spring profiles. Using the example above, you could just create application-prod.yml with a "prod" specific value for poolMaxTotal. Then start your app with --spring.profiles.active=prod and the "prod" value will be used instead of your default value in application.yml. You can do this for however many environments you need.
application-prod.yml
poolMaxTotal: 20
For an async HttpClient, see here: http://vincentdevillers.blogspot.fr/2013/10/a-best-spring-asyncresttemplate.html

Related

setting timeout using #ConfigurationProperties in SpringBoot

I am going through a code that configures dedicated restTemplate for a rest operation. I see the following properties
httpProperties.connection-request-timeout=6100
httpProperties.connect-timeout=6100
httpProperties.read-timeout=6100
My Config class looks like below
#Bean
#ConfigurationProperties(prefix = "httpProperties")
public HttpComponentsClientHttpRequestFactory webSystemHttpRequestFactory() {
SSLContext sslContext;
try {
SSLConnectionSocketFactory socketFactory = new SSLConnectionSocketFactory(sslContext);
HttpClient httpClient = HttpClients.custom().setMaxConnTotal(maxTotalConnection)
.setMaxConnPerRoute(maxConnectionPerRoute).setSSLSocketFactory(socketFactory).build();
return new HttpComponentsClientHttpRequestFactory(httpClient);
}
catch(Exception e) {
}
return new HttpComponentsClientHttpRequestFactory();
}
#Bean(name = "webSystemRestTemplate")
public RestTemplate webSystemRestTemplate() {
RestTemplate restTemplate = new RestTemplate(webSystemHttpRequestFactory());
return restTemplate;
}
I can see the logs
o.a.h.i.c.DefaultManagedHttpClientConnection.setSocketTimeout - http-outgoing-1: set socket timeout to 6100
Here is what i want to understand:
How is this value set and to which property by the #CnfigurationProperties annotation?
Is it applicable at the spring boot application level or at each request level?
Please help me understand the concept underlying.
Note: Apache http client version used is 4.5.2
In the source code for HttpComponentsClientHttpRequestFactory.class there is an object called RequestConfig.class.
In it's source code you can see that there are three parameters.
private final Timeout connectionRequestTimeout;
private final Timeout connectTimeout;
private final Timeout responseTimeout;
These are the ones that the parameters map to using
#ConfigurationProperties(prefix = "httpProperties")
That is not the most common way to set these parameters. But there are multiple ways to set these as pointed out here.
RestTemplate timeout examples
The properties are setting the attributes connectionRequestTimeOut, connectTimeOut and readTimeOut of the HttpComponentsClientHttpRequestFactory class. The mapping is done using the ConfigurationProperties annotation that maps the kebab case property names to the bean attributes.
HttpComponentsClientHttpRequestFactory documentation :
HttpComponentsClientHttpRequestFactory

Spring cloud multiple RestTemplate

I am using spring cloud: Spring Boot Application with Eureka + Ribbon default configuration.
I am using 2 RestTemplate configurations, both are #LoadBalanced currently and both of them have the same UriTemplateHandler.
I declared both the #SpringBootApplication and also the #RibbonClient(name="${service.name}") annotations.
My problem is:
When I am trying to access the first configured RestTemplate, the RestTemplate resolvs (by eureka and load balancing by ribbon) to a server , not as I requested as configured in the UriTemplateHandler.
For example: in the UriTemplateHandler I configured "A-Service" and in real time the restTemplate sends the httpRequest to "B-Service"
This behavior happens often, not just for a specific request, but it looks like it only happens when I'm accessing the first configured RestTemplate.
Is it a problem to use 2 RestTemplate with the same uri?
I have no idea why it happens, please advise.
When creating these rest templates as beans, name them uniquely, like e.g.
#LoadBalanced
#Bean("integrationRestTemplate")
public RestTemplate restTemplate() throws Exception {
// build and return your rest template
return ....
}
Then, the other one might be without any specific name e.g.
#Bean
public RestTemplate restTemplate() {
return new RestTemplate();
}
Now, if you have these two distinctive rest templates, you can inject the former one e.g. like that:
#Service
public class MyService {
private final RestTemplate restTemplate;
public ApplicantService(#Qualifier("integrationRestTemplate") RestTemplate restTemplate) {
this.restTemplate = restTemplate;
}
// service methods
...
}
Basically, the point is you can choose whatever rest template you want, by specifying a #Qualifier.

spring boot 1.3.3 create multiple resttemplate per env

I am on spring boot 1.3.3 version, I have a requirement where my spring boot application need to call endpoint(s) based on env passed,
which means if env passed as Dev i would need to call devendpoint,
if env passed as Dev1 then need to call dev1endpoint and so on.
So how can I do this ?
Do I need to create multiple restTemplate instances ?
Should I construct the resttemplate dynamically based on env passed ?
As part of constructing resttemplate i would also need to add appllicable interceptor based on env selected.
Plesae suggest.
You can have two beans of the same class. One can be labeled as the primary, and the use on the #Autowired can specify which one to use with the #Qualifier.
Example:
#Configuration
public class MyConfig {
#Bean
#Primary
public RestTemplate typicalConfig() {
// various configs on your rest template
return new RestTemplate();
}
#Bean
public RestTemplate lessTypical() {
// various alternate configurations
return new RestTemplate();
}
}
Now in your service class:
#Service
public class MyService {
#Autowired
RestTemplate typicalRestTemplate;
#Autowired
#Qualifier("lessTypical")
private RestTemplate alternateRestTemplate;
...
}
correct me if I didn't understand your question. I understand that you are going to have different environments but you are going to change this endpoints in runtime depends on some information or whatever, but I don't understand the part when you said you have to create multiple instances of restTemplate for that environments, from my experience on spring boot applications you don't have to do things like that, You just have to create your restTemplate configuration bean.
#Bean
public RestTemplate restTemplate() {
return new RestTemplate();
}
And then injected that object on your services class and do whatever you want with them. I recommend you to read the follow article about restTempalte may this could help you http://www.baeldung.com/rest-template

How to overwrite Spring Cloud OAuth2 client autoconfiguration?

We want to setup a microservice which provides a REST API so it is configured as a OAuth2 resource server. This service should also act as a OAuth2 client with the client credential grant. Here is the configuration:
spring.oauth2.client.id=clientCredentialsResource
spring.oauth2.client.accessTokenUri=http://localhost:9003/oauth/token
spring.oauth2.client.userAuthorizationUri=http://localhost:9003/oauth/authorize
spring.oauth2.client.grantType=client_credentials
spring.oauth2.client.clientId=<service-id>
spring.oauth2.client.clientSecret=<service-pw>
The resource server part works fine. For the client part we want to use Feign, Ribbon and Eureka:
#FeignClient("user")
public interface UserClient
{
#RequestMapping( method = RequestMethod.GET, value = "/user/{uid}")
Map<String, String> getUser(#PathVariable("uid") String uid);
}
Based on the gist in issue https://github.com/spring-cloud/spring-cloud-security/issues/56 I created a feign request intercepter which sets the access token from the autowired OAuth2RestOperations template in the feign request header
#Autowired
private OAuth2RestOperations restTemplate;
template.header(headerName, String.format("%s %s", tokenTypeName, restTemplate.getAccessToken().toString()));
But this gives me the error on calling the user service:
error="access_denied", error_description="Unable to obtain a new access token for resource 'clientCredentialsResource'. The provider manager is not configured to support it.
As I can see the OAuth2ClientAutoConfiguration creates always an instance of AuthorizationCodeResourceDetails for an web application but not the required ClientCredentialsResourceDetails which is only used for non-web applications. In the end the no access token privider is responsible for the resource details and the call failed in
AccessTokenProviderChain.obtainNewAccessTokenInternal(AccessTokenProviderChain.java:146)
I tried to overwrite the auto configuration but failed. Can somebody please give me a hint how to do it?
To switch off this piece of autoconfiguration you can set spring.oauth2.client.clientId= (empty), (per the source code), otherwise you have to "exclude" it in the #EnableAutoConfiguration. If you do that you can just set up your own OAuth2RestTemplate and fill in the "real" client ID from your own configuration, e.g.
#Configuration
#EnableOAuth2Client
public class MyConfiguration {
#Value("myClientId")
String myClientId;
#Bean
#ConfigurationProperties("spring.oauth2.client")
#Primary
public ClientCredentialsResourceDetails oauth2RemoteResource() {
ClientCredentialsResourceDetails details = new ClientCredentialsResourceDetails();
details.setClientId(myClientId);
return details;
}
#Bean
public OAuth2ClientContext oauth2ClientContext() {
return new DefaultOAuth2ClientContext(new DefaultAccessTokenRequest());
}
#Bean
#Primary
public OAuth2RestTemplate oauth2RestTemplate(
OAuth2ClientContext oauth2ClientContext,
OAuth2ProtectedResourceDetails details) {
OAuth2RestTemplate template = new OAuth2RestTemplate(details,
oauth2ClientContext);
return template;
}
}

Spring RestTemplate timeout

I would like to set the connection timeouts for a rest service used by my web application. I'm using Spring's RestTemplate to talk to my service. I've done some research and I've found and used the xml below (in my application xml) which I believe is meant to set the timeout. I'm using Spring 3.0.
I've also seen the same problem here Timeout configuration for spring webservices with RestTemplate but the solutions don't seem that clean, I'd prefer to set the timeout values via Spring config
<bean id="RestOperations" class="org.springframework.web.client.RestTemplate">
<constructor-arg>
<bean class="org.springframework.http.client.CommonsClientHttpRequestFactory">
<property name="readTimeout" value="${restURL.connectionTimeout}" />
</bean>
</constructor-arg>
</bean>
It seems whatever I set the readTimeout to be I get the following:
Network cable disconnected:
Waits about 20 seconds and reports following exception:
org.springframework.web.client.ResourceAccessException: I/O error: No route to host: connect; nested exception is java.net.NoRouteToHostException: No route to host: connect
Url incorrect so 404 returned by rest service:
Waits about 10 seconds and reports following exception:
org.springframework.web.client.HttpClientErrorException: 404 Not Found
My requirements require shorter timeouts so I need to be able to change these. Any ideas as to what I'm doing wrong?
Many thanks.
For Spring Boot >= 1.4
#Configuration
public class AppConfig
{
#Bean
public RestTemplate restTemplate(RestTemplateBuilder restTemplateBuilder)
{
return restTemplateBuilder
.setConnectTimeout(...)
.setReadTimeout(...)
.build();
}
}
For Spring Boot <= 1.3
#Configuration
public class AppConfig
{
#Bean
#ConfigurationProperties(prefix = "custom.rest.connection")
public HttpComponentsClientHttpRequestFactory customHttpRequestFactory()
{
return new HttpComponentsClientHttpRequestFactory();
}
#Bean
public RestTemplate customRestTemplate()
{
return new RestTemplate(customHttpRequestFactory());
}
}
then in your application.properties
custom.rest.connection.connection-request-timeout=...
custom.rest.connection.connect-timeout=...
custom.rest.connection.read-timeout=...
This works because HttpComponentsClientHttpRequestFactory has public setters connectionRequestTimeout, connectTimeout, and readTimeout and #ConfigurationProperties sets them for you.
For Spring 4.1 or Spring 5 without Spring Boot using #Configuration instead of XML
#Configuration
public class AppConfig
{
#Bean
public RestTemplate customRestTemplate()
{
HttpComponentsClientHttpRequestFactory httpRequestFactory = new HttpComponentsClientHttpRequestFactory();
httpRequestFactory.setConnectionRequestTimeout(...);
httpRequestFactory.setConnectTimeout(...);
httpRequestFactory.setReadTimeout(...);
return new RestTemplate(httpRequestFactory);
}
}
I finally got this working.
I think the fact that our project had two different versions of the commons-httpclient jar wasn't helping. Once I sorted that out I found you can do two things...
In code you can put the following:
HttpComponentsClientHttpRequestFactory rf =
(HttpComponentsClientHttpRequestFactory) restTemplate.getRequestFactory();
rf.setReadTimeout(1 * 1000);
rf.setConnectTimeout(1 * 1000);
The first time this code is called it will set the timeout for the HttpComponentsClientHttpRequestFactory class used by the RestTemplate. Therefore, all subsequent calls made by RestTemplate will use the timeout settings defined above.
Or the better option is to do this:
<bean id="RestOperations" class="org.springframework.web.client.RestTemplate">
<constructor-arg>
<bean class="org.springframework.http.client.HttpComponentsClientHttpRequestFactory">
<property name="readTimeout" value="${application.urlReadTimeout}" />
<property name="connectTimeout" value="${application.urlConnectionTimeout}" />
</bean>
</constructor-arg>
</bean>
Where I use the RestOperations interface in my code and get the timeout values from a properties file.
This question is the first link for a Spring Boot search, therefore, would be great to put here the solution recommended in the official documentation. Spring Boot has its own convenience bean RestTemplateBuilder:
#Bean
public RestTemplate restTemplate(
RestTemplateBuilder restTemplateBuilder) {
return restTemplateBuilder
.setConnectTimeout(Duration.ofSeconds(500))
.setReadTimeout(Duration.ofSeconds(500))
.build();
}
Manual creation of RestTemplate instances is a potentially troublesome approach because other auto-configured beans are not being injected in manually created instances.
Here are my 2 cents. Nothing new, but some explanations, improvements and newer code.
By default, RestTemplate has infinite timeout.
There are two kinds of timeouts: connection timeout and read time out. For instance, I could connect to the server but I could not read data. The application was hanging and you have no clue what's going on.
I am going to use annotations, which these days are preferred over XML.
#Configuration
public class AppConfig {
#Bean
public RestTemplate restTemplate() {
var factory = new SimpleClientHttpRequestFactory();
factory.setConnectTimeout(3000);
factory.setReadTimeout(3000);
return new RestTemplate(factory);
}
}
Here we use SimpleClientHttpRequestFactory to set the connection and read time outs.
It is then passed to the constructor of RestTemplate.
#Configuration
public class AppConfig {
#Bean
public RestTemplate restTemplate(RestTemplateBuilder builder) {
return builder
.setConnectTimeout(Duration.ofMillis(3000))
.setReadTimeout(Duration.ofMillis(3000))
.build();
}
}
In the second solution, we use the RestTemplateBuilder. Also notice the parameters of the two methods: they take Duration. The overloaded methods that take directly milliseconds are now deprecated.
Edit
Tested with Spring Boot 2.1.0 and Java 11.
Here is a really simple way to set the timeout:
RestTemplate restTemplate = new RestTemplate(getClientHttpRequestFactory());
private ClientHttpRequestFactory getClientHttpRequestFactory() {
int timeout = 5000;
HttpComponentsClientHttpRequestFactory clientHttpRequestFactory =
new HttpComponentsClientHttpRequestFactory();
clientHttpRequestFactory.setConnectTimeout(timeout);
return clientHttpRequestFactory;
}
RestTemplate timeout with SimpleClientHttpRequestFactory
To programmatically override the timeout properties, we can customize the SimpleClientHttpRequestFactory class as below.
Override timeout with SimpleClientHttpRequestFactory
//Create resttemplate
RestTemplate restTemplate = new RestTemplate(getClientHttpRequestFactory());
//Override timeouts in request factory
private SimpleClientHttpRequestFactory getClientHttpRequestFactory()
{
SimpleClientHttpRequestFactory clientHttpRequestFactory
= new SimpleClientHttpRequestFactory();
//Connect timeout
clientHttpRequestFactory.setConnectTimeout(10_000);
//Read timeout
clientHttpRequestFactory.setReadTimeout(10_000);
return clientHttpRequestFactory;
}
RestTemplate timeout with HttpComponentsClientHttpRequestFactory
SimpleClientHttpRequestFactory helps in setting timeout but it is very limited in functionality and may not prove sufficient in realtime applications. In production code, we may want to use HttpComponentsClientHttpRequestFactory which support HTTP Client library along with resttemplate.
HTTPClient provides other useful features such as connection pool, idle connection management etc.
Read More : Spring RestTemplate + HttpClient configuration example
Override timeout with HttpComponentsClientHttpRequestFactory
//Create resttemplate
RestTemplate restTemplate = new RestTemplate(getClientHttpRequestFactory());
//Override timeouts in request factory
private SimpleClientHttpRequestFactory getClientHttpRequestFactory()
{
HttpComponentsClientHttpRequestFactory clientHttpRequestFactory
= new HttpComponentsClientHttpRequestFactory();
//Connect timeout
clientHttpRequestFactory.setConnectTimeout(10_000);
//Read timeout
clientHttpRequestFactory.setReadTimeout(10_000);
return clientHttpRequestFactory;
}
reference: Spring RestTemplate timeout configuration example
Simple timeout for restTemplate.I have set the read and write timeout for 3 seconds.
#Bean
public RestTemplate restTemplate(RestTemplateBuilder restTemplateBuilder){
RestTemplate restTemplate= restTemplateBuilder.setConnectTimeout(Duration.ofMillis(3000)).setReadTimeout(Duration.ofMillis(3000)).build();
return restTemplate;
}
I had a similar scenario, but was also required to set a Proxy. The simplest way I could see to do this was to extend the SimpleClientHttpRequestFactory for the ease of setting the proxy (different proxies for non-prod vs prod). This should still work even if you don't require the proxy though. Then in my extended class I override the openConnection(URL url, Proxy proxy) method, using the same as the source, but just setting the timeouts before returning.
#Override
protected HttpURLConnection openConnection(URL url, Proxy proxy) throws IOException {
URLConnection urlConnection = proxy != null ? url.openConnection(proxy) : url.openConnection();
Assert.isInstanceOf(HttpURLConnection.class, urlConnection);
urlConnection.setConnectTimeout(5000);
urlConnection.setReadTimeout(5000);
return (HttpURLConnection) urlConnection;
}
To expand on benscabbia's answer:
private RestTemplate restCaller = new RestTemplate(getClientHttpRequestFactory());
private ClientHttpRequestFactory getClientHttpRequestFactory() {
int connectionTimeout = 5000; // milliseconds
int socketTimeout = 10000; // milliseconds
RequestConfig config = RequestConfig.custom()
.setConnectTimeout(connectionTimeout)
.setConnectionRequestTimeout(connectionTimeout)
.setSocketTimeout(socketTimeout)
.build();
CloseableHttpClient client = HttpClientBuilder
.create()
.setDefaultRequestConfig(config)
.build();
return new HttpComponentsClientHttpRequestFactory(client);
}
private static RestTemplate restTemplate;
static {
HttpComponentsClientHttpRequestFactory rf = new HttpComponentsClientHttpRequestFactory();
rf.setReadTimeout(3 * 1000);
rf.setConnectTimeout(2 * 1000);
restTemplate = new RestTemplate(rf);
restTemplate.getMessageConverters()
.add(0, new StringHttpMessageConverter(StandardCharsets.UTF_8));
}
Setting the timeout only in RestTemplateBuilder didn't work for me when i was using Apache's httpcomponents. I had to set the timeout in the RequestFactory as well.
Here's the entire code:
public RestTemplate templateBuilder() {
RestTemplate restTemplate = this.restTemplateBuilder
.setConnectTimeout(Duration.ofSeconds(connectTimeout))
.setReadTimeout(Duration.ofSeconds(readTimeout))
.build();
HttpComponentsClientHttpRequestFactory requestFactory = new HttpComponentsClientHttpRequestFactory();
requestFactory.setConnectTimeout((int) connectTimeout * 1000);
requestFactory.setReadTimeout((int) readTimeout * 1000);
requestFactory.setConnectionRequestTimeout((int) connectTimeout * 1000);
restTemplate.setRequestFactory(new BufferingClientHttpRequestFactory(requestFactory));
return restTemplate;
}

Resources