How to cache response with Jersey?

Question

I am trying to develop restful API using Jersey.I have GET API's for a particular get operation my GET is taking same time from the same client. Is it possible to cache response? Any pointers is appreciated.

Thanks

Answer 1

You can use CacheControl, eTag - follow below example code

// In your jersey method
    final EntityTag eTag = new EntityTag(resource.getId() + "_" +
     resource.getLastModified().getTime());
    final CacheControl cacheControl = new CacheControl();
    cacheControl.setMaxAge(-1);

    ResponseBuilder builder = request.evaluatePreconditions(
         resource.getLastModified(), eTag);

    // the resoruce's information was modified, return it
    if (builder == null) {
         builder = Response.ok(resource);
    }

    // the resource's information was not modified, return a 304

    return builder.cacheControl(cacheControl).lastModified(
         resource.getLastModified()).tag(eTag).build();

Replace resource with your Resource instance.

Answer 2

Summary of solutions:

1. Request as method parameter

   Interface:

    @Path("myentity") public interface MyEntityResource @GET @Produces(MediaType.APPLICATION_JSON) public Response getMyEntity(@Context final Request request); }

   Implementation:

    public class MyEntityResourceImpl implements MyEntityResource @Override public Response getMyEntity(final Request request) { final MyEntity myEntity = ... // load entity final String eTagValue = ... // calclutate value of ETag final EntityTag eTag = new EntityTag(eTagValue); ResponseBuilder responseBuilder = request.evaluatePreconditions(eTag); if (responseBuilder == null) { return Response.ok(user).tag(eTag).build(); } return responseBuilder.build(); } }

   Disadvantages:

   • implementation detail Request is exposed

   • return type Reponse is generic

   • missing grammar of return type in WADL

   • client proxy with unnecessary parameter Request

2. Request as instance variable

   Interface:

    @Path("myentity") public interface MyEntityResource @GET @Produces(MediaType.APPLICATION_JSON) public Response getMyEntity(); }

   Implementation:

    public class MyEntityResourceImpl implements MyEntityResource @Context private Request request @Override public Response getMyEntity() { final MyEntity myEntity = ... // load entity final String eTagValue = ... // calclutate value of ETag final EntityTag eTag = new EntityTag(eTagValue); ResponseBuilder responseBuilder = request.evaluatePreconditions(eTag); if (responseBuilder == null) { return Response.ok(user).tag(eTag).build(); } return responseBuilder.build(); } }

   Disadvantages:

   • return type Reponse is generic

   • missing grammar of return type in WADL

   • dependency injection with @Context is complicated, see https://stackoverflow.com/questions/33240443

3. ShallowEtagHeaderFilter as web filter

   web.xml:

    <filter> <filter-name>etagFilter</filter-name> <filter-class>org.springframework.web.filter.ShallowEtagHeaderFilter</filter-class> </filter> <filter-mapping> <filter-name>etagFilter</filter-name> <url-pattern>/api/*</url-pattern> </filter-mapping>

   Interface:

    @Path("myentity") public interface MyEntityResource @GET @Produces(MediaType.APPLICATION_JSON) public MyEntity getMyEntity(); }

   Implementation:

    public class MyEntityResourceImpl implements MyEntityResource @Override public MyEntity getMyEntity() { final MyEntity myEntity = ... // load entity return myEntity; } }

   Disadvantages:

   • bad server performance, see JavaDoc

   • works only on uncommitted response

   • no support of weak ETag

4. Custom WriterInterceptor as JAX-RS Interceptor

   Interceptor:

    public class CustomInterceptor implements WriterInterceptor { @Context private Request request; @Override public void aroundWriteTo(WriterInterceptorContext context) throws IOException, WebApplicationException { OutputStream old = context.getOutputStream(); ByteArrayOutputStream buffer = new ByteArrayOutputStream(); try { context.setOutputStream(buffer); context.proceed(); byte[] entity = buffer.toByteArray(); String etag = ... // calclutate value of ETag context.getHeaders().putSingle(HttpHeaders.ETAG, etag); ResponseBuilder responseBuilder = request.evaluatePreconditions(eTag); if (responseBuilder == null) { throw new WebApplicationException(responseBuilder.status(Response.Status.NOT_MODIFIED).header(HttpHeaders.ETAG, etag).build()); } old.write(entity); } finally { context.setOutputStream(old); } } }

   See also: ServerCacheInterceptor (Resteasy)

   Interface:

    @Path("myentity") public interface MyEntityResource @GET @Produces(MediaType.APPLICATION_JSON) public MyEntity getMyEntity(); }

   Implementation:

    public class MyEntityResourceImpl implements MyEntityResource @Override public MyEntity getMyEntity() { final MyEntity myEntity = ... // load entity return myEntity; } }

   Disadvantages:

   • no predefined interceptor for Jersey available

   • bad server performance

   • no support of weak ETag

   • ugly workaround with WebApplicationException

Answer 3

You could use any caching mechanism applicable for standard java together with Jersey, like Ehcache .

You only have to pay attention to verify that your data in the backend hasn't changed.

Here is simple example with Ehcache :

@GET
@Path("{id}")
public List<Data> getData(@PathParam("id") Long id) {
    Element element = CacheManager.getInstance().getCache("test").get(id);
    if(element == null) {
        Data value = fetchElementFromBackend(id);
        CacheManager.getInstance().getCache("test").put(new Element(id, value));
        return value;
    }

    return element.getObjectValue();
}

Answer 4

Recently I've been solving a similar (if not the same) problem. As a side result of it the following library emerged: https://github.com/AndreyLebedenko/dropwizard-caching-filter

The way it can be used is like below:

@GET
@Produces(MediaType.APPLICATION_JSON)
@Path("/cached")
@ResponseCachedByFilter(10000)
public Object getCached() {
    return dao.get();
} 

Hope it helps.