The English version of quarkus.io is the official project site. Translated sites are community supported on a best-effort basis.

Application Data Caching

In this guide, you will learn how to enable application data caching in any CDI managed bean of your Quarkus application.

这项技术被认为是preview。

preview(预览) 中,不保证向后兼容和在生态系统中的存在。具体的改进可能需要改变配置或API,并且正在计划变得 稳定 。欢迎在我们的 邮件列表 中提供反馈,或在我们的 GitHub问题列表 中提出问题。

For a full list of possible statuses, check our FAQ entry.

Prerequisites

要完成这个指南,你需要:

  • 大概15分钟

  • 编辑器

  • 安装JDK 11以上版本并正确配置了 JAVA_HOME

  • Apache Maven 3.8.1+

  • 如果你愿意的话,还可以选择使用Quarkus CLI

  • 如果你想构建原生可执行程序,可以选择安装Mandrel或者GraalVM,并正确配置(或者使用Docker在容器中进行构建)

Scenario

Let’s imagine you want to expose in your Quarkus application a REST API that allows users to retrieve the weather forecast for the next three days. The problem is that you have to rely on an external meteorological service which only accepts requests for one day at a time and takes forever to answer. Since the weather forecast is updated once every twelve hours, caching the service responses would definitely improve your API performances.

We’ll do that using a single Quarkus annotation.

Solution

We recommend that you follow the instructions in the next sections and create the application step by step. However, you can go right to the completed example.

Clone the Git repository: git clone https://github.com/quarkusio/quarkus-quickstarts.git, or download an archive.

The solution is located in the cache-quickstart directory.

Creating the Maven project

First, we need to create a new Quarkus project with the following command:

CLI
quarkus create app org.acme:cache-quickstart \
    --extension=cache,resteasy-reactive-jackson \
    --no-code
cd cache-quickstart

创建Grade项目,请添加 --gradle 或者 --gradle-kotlin-dsl 参数。

关于如何安装并使用Quarkus CLI的更多信息,请参考Quarkus CLI指南

Maven
mvn io.quarkus.platform:quarkus-maven-plugin:2.11.2.Final:create \
    -DprojectGroupId=org.acme \
    -DprojectArtifactId=cache-quickstart \
    -Dextensions="cache,resteasy-reactive-jackson" \
    -DnoCode
cd cache-quickstart

创建Grade项目,请添加 -DbuildTool=gradle 或者 -DbuildTool=gradle-kotlin-dsl 参数。

This command generates the project and imports the cache and resteasy-reactive-jackson extensions.

If you already have your Quarkus project configured, you can add the cache extension to your project by running the following command in your project base directory:

CLI
quarkus extension add 'cache'
Maven
./mvnw quarkus:add-extension -Dextensions="cache"
Gradle
./gradlew addExtension --extensions="cache"

This will add the following to your build file:

pom.xml
<dependency>
    <groupId>io.quarkus</groupId>
    <artifactId>quarkus-cache</artifactId>
</dependency>
build.gradle
implementation("io.quarkus:quarkus-cache")

Creating the REST API

Let’s start by creating a service that will simulate an extremely slow call to the external meteorological service. Create src/main/java/org/acme/cache/WeatherForecastService.java with the following content:

package org.acme.cache;

import java.time.LocalDate;

import javax.enterprise.context.ApplicationScoped;

@ApplicationScoped
public class WeatherForecastService {

    public String getDailyForecast(LocalDate date, String city) {
        try {
            Thread.sleep(2000L); (1)
        } catch (InterruptedException e) {
            Thread.currentThread().interrupt();
        }
        return date.getDayOfWeek() + " will be " + getDailyResult(date.getDayOfMonth() % 4) + " in " + city;
    }

    private String getDailyResult(int dayOfMonthModuloFour) {
        switch (dayOfMonthModuloFour) {
            case 0:
                return "sunny";
            case 1:
                return "cloudy";
            case 2:
                return "chilly";
            case 3:
                return "rainy";
            default:
                throw new IllegalArgumentException();
        }
    }
}
1 This is where the slowness comes from.

We also need a class that will contain the response sent to the users when they ask for the next three days weather forecast. Create src/main/java/org/acme/cache/WeatherForecast.java this way:

package org.acme.cache;

import java.util.List;

public class WeatherForecast {

    private List<String> dailyForecasts;

    private long executionTimeInMs;

    public WeatherForecast(List<String> dailyForecasts, long executionTimeInMs) {
        this.dailyForecasts = dailyForecasts;
        this.executionTimeInMs = executionTimeInMs;
    }

    public List<String> getDailyForecasts() {
        return dailyForecasts;
    }

    public long getExecutionTimeInMs() {
        return executionTimeInMs;
    }
}

Now, we just need to create the REST resource. Create the src/main/java/org/acme/cache/WeatherForecastResource.java file with this content:

package org.acme.cache;

import java.time.LocalDate;
import java.util.Arrays;
import java.util.List;

import javax.inject.Inject;
import javax.ws.rs.GET;
import javax.ws.rs.Path;
import javax.ws.rs.core.MediaType;

import org.jboss.resteasy.reactive.RestQuery;

@Path("/weather")
public class WeatherForecastResource {

    @Inject
    WeatherForecastService service;

    @GET
    public WeatherForecast getForecast(@RestQuery String city, @RestQuery long daysInFuture) { (1)
        long executionStart = System.currentTimeMillis();
        List<String> dailyForecasts = Arrays.asList(
                service.getDailyForecast(LocalDate.now().plusDays(daysInFuture), city),
                service.getDailyForecast(LocalDate.now().plusDays(daysInFuture + 1L), city),
                service.getDailyForecast(LocalDate.now().plusDays(daysInFuture + 2L), city));
        long executionEnd = System.currentTimeMillis();
        return new WeatherForecast(dailyForecasts, executionEnd - executionStart);
    }
}
1 If the daysInFuture query parameter is omitted, the three days weather forecast will start from the current day. Otherwise, it will start from the current day plus the daysInFuture value.

We’re all done! Let’s check if everything’s working.

First, run the application using dev mode from the project directory:

CLI
quarkus dev
Maven
./mvnw quarkus:dev
Gradle
./gradlew --console=plain quarkusDev

Then, call http://localhost:8080/weather?city=Raleigh from a browser. After six long seconds, the application will answer something like this:

{"dailyForecasts":["MONDAY will be cloudy in Raleigh","TUESDAY will be chilly in Raleigh","WEDNESDAY will be rainy in Raleigh"],"executionTimeInMs":6001}

The response content may vary depending on the day you run the code.

You can try calling the same URL again and again, it will always take six seconds to answer.

Enabling the cache

Now that your Quarkus application is up and running, let’s tremendously improve its response time by caching the external meteorological service responses. Update the WeatherForecastService class like this:

package org.acme.cache;

import java.time.LocalDate;

import javax.enterprise.context.ApplicationScoped;

import io.quarkus.cache.CacheResult;

@ApplicationScoped
public class WeatherForecastService {

    @CacheResult(cacheName = "weather-cache") (1)
    public String getDailyForecast(LocalDate date, String city) {
        try {
            Thread.sleep(2000L);
        } catch (InterruptedException e) {
            Thread.currentThread().interrupt();
        }
        return date.getDayOfWeek() + " will be " + getDailyResult(date.getDayOfMonth() % 4) + " in " + city;
    }

    private String getDailyResult(int dayOfMonthModuloFour) {
        switch (dayOfMonthModuloFour) {
            case 0:
                return "sunny";
            case 1:
                return "cloudy";
            case 2:
                return "chilly";
            case 3:
                return "rainy";
            default:
                throw new IllegalArgumentException();
        }
    }
}
1 We only added this annotation (and the associated import of course).

Let’s try to call http://localhost:8080/weather?city=Raleigh again. You’re still waiting a long time before receiving an answer. This is normal since the server just restarted and the cache was empty.

Wait a second! The server restarted by itself after the WeatherForecastService update? Yes, this is one of Quarkus amazing features for developers called live coding.

Now that the cache was loaded during the previous call, try calling the same URL. This time, you should get a super fast answer with an executionTimeInMs value close to 0.

Let’s see what happens if we start from one day in the future using the http://localhost:8080/weather?city=Raleigh&daysInFuture=1 URL. You should get an answer two seconds later since two of the requested days were already loaded in the cache.

You can also try calling the same URL with a different city and see the cache in action again. The first call will take six seconds and the following ones will be answered immediately.

Congratulations! You just added application data caching to your Quarkus application with a single line of code!

Do you want to learn more about the Quarkus application data caching abilities? The following sections will show you everything there is to know about it.

Caching using annotations

Quarkus offers a set of annotations that can be used in a CDI managed bean to enable caching abilities.

Caching annotations are not allowed on private methods. They will work fine with any other access modifier including package-private (no explicit modifier).

@CacheResult

Loads a method result from the cache without executing the method body whenever possible.

When a method annotated with @CacheResult is invoked, Quarkus will compute a cache key and use it to check in the cache whether the method has been already invoked. See the Cache keys building logic section of this guide to learn how the cache key is computed. If a value is found in the cache, it is returned and the annotated method is never actually executed. If no value is found, the annotated method is invoked and the returned value is stored in the cache using the computed key.

A method annotated with CacheResult is protected by a lock on cache miss mechanism. If several concurrent invocations try to retrieve a cache value from the same missing key, the method will only be invoked once. The first concurrent invocation will trigger the method invocation while the subsequent concurrent invocations will wait for the end of the method invocation to get the cached result. The lockTimeout parameter can be used to interrupt the lock after a given delay. The lock timeout is disabled by default, meaning the lock is never interrupted. See the parameter Javadoc for more details.

This annotation cannot be used on a method returning void.

Quarkus is able to also cache null values unlike the underlying Caffeine provider. See more on this topic below.

@CacheInvalidate

Removes an entry from the cache.

When a method annotated with @CacheInvalidate is invoked, Quarkus will compute a cache key and use it to try to remove an existing entry from the cache. See the Cache keys building logic section of this guide to learn how the cache key is computed. If the key does not identify any cache entry, nothing will happen.

@CacheInvalidateAll

When a method annotated with @CacheInvalidateAll is invoked, Quarkus will remove all entries from the cache.

@CacheKey

When a method argument is annotated with @CacheKey, it is identified as a part of the cache key during an invocation of a method annotated with @CacheResult or @CacheInvalidate.

This annotation is optional and should only be used when some method arguments are NOT part of the cache key.

Cache keys building logic

Cache keys are built by the annotations API using the following logic:

  • If an io.quarkus.cache.CacheKeyGenerator is declared in a @CacheResult or a @CacheInvalidate annotation, then it is used to generate the cache key. The @CacheKey annotations that might be present on some method arguments are ignored.

  • Otherwise, if the method has no arguments, then the cache key is an instance of io.quarkus.cache.DefaultCacheKey built from the cache name.

  • Otherwise, if the method has exactly one argument, then that argument is the cache key.

  • Otherwise, if the method has multiple arguments but only one annotated with @CacheKey, then that annotated argument is the cache key.

  • Otherwise, if the method has multiple arguments annotated with @CacheKey, then the cache key is an instance of io.quarkus.cache.CompositeCacheKey built from these annotated arguments.

  • Otherwise, if the method has multiple arguments and none of them are annotated with @CacheKey, the cache key is an instance of io.quarkus.cache.CompositeCacheKey built from all the method arguments.

Each non-primitive method argument that is part of the key must implement equals() and hashCode() correctly for the cache to work as expected.

When a cache key is built from several method arguments, whether they are explicitly identified with @CacheKey or not, the building logic depends on the order of these arguments in the method signature. On the other hand, the arguments names are not used at all and do not have any effect on the cache key.

package org.acme.cache;

import javax.enterprise.context.ApplicationScoped;

import io.quarkus.cache.CacheInvalidate;
import io.quarkus.cache.CacheResult;

@ApplicationScoped
public class CachedService {

    @CacheResult(cacheName = "foo")
    public Object load(String keyElement1, Integer keyElement2) {
        // Call expensive service here.
    }

    @CacheInvalidate(cacheName = "foo")
    public void invalidate1(String keyElement2, Integer keyElement1) { (1)
    }

    @CacheInvalidate(cacheName = "foo")
    public void invalidate2(Integer keyElement2, String keyElement1) { (2)
    }

    @CacheInvalidate(cacheName = "foo")
    public void invalidate3(Object notPartOfTheKey, @CacheKey String keyElement1, @CacheKey Integer keyElement2) { (3)
    }

    @CacheInvalidate(cacheName = "foo")
    public void invalidate4(Object notPartOfTheKey, @CacheKey Integer keyElement2, @CacheKey String keyElement1) { (4)
    }
}
1 Calling this method WILL invalidate values cached by the load method even if the key elements names have been swapped.
2 Calling this method WILL NOT invalidate values cached by the load method because the key elements order is different.
3 Calling this method WILL invalidate values cached by the load method because the key elements order is the same.
4 Calling this method WILL NOT invalidate values cached by the load method because the key elements order is different.

Generating a cache key with CacheKeyGenerator

You may want to include more than the arguments of a method into a cache key. This can be done by implementing the io.quarkus.cache.CacheKeyGenerator interface and declaring that implementation in the keyGenerator field of a @CacheResult or @CacheInvalidate annotation.

If a CDI scope is declared on a key generator class and if that class has a default qualifier (no qualifier annotation or @javax.enterprise.inject.Default), then the key generator will be injected as a CDI bean during the cache key computation. Otherwise, the key generator will be instantiated using its default constructor. All CDI scopes supported by Quarkus can be used on a key generator.

The following key generator will be injected as a CDI bean:

package org.acme.cache;

import java.lang.reflect.Method;
import javax.enterprise.context.ApplicationScoped;
import javax.inject.Inject;

import io.quarkus.cache.CacheKeyGenerator;
import io.quarkus.cache.CompositeCacheKey;

@ApplicationScoped
public class ApplicationScopedKeyGen implements CacheKeyGenerator {

    @Inject
    AnythingYouNeedHere anythingYouNeedHere; (1)

    @Override
    public Object generate(Method method, Object... methodParams) { (2)
        return new CompositeCacheKey(anythingYouNeedHere.getData(), methodParams[1]); (3)
    }
}
1 External data can be included into the cache key by injecting a CDI bean in the key generator.
2 Be careful while using Method, some of its methods can be expensive.
3 Make sure the method has enough arguments before accessing them from their index. Otherwise, an IndexOutOfBoundsException may be thrown during the cache key computation.

The following key generator will be instantiated using its default constructor:

package org.acme.cache;

import java.lang.reflect.Method;

import io.quarkus.cache.CacheKeyGenerator;
import io.quarkus.cache.CompositeCacheKey;

public class NotABeanKeyGen implements CacheKeyGenerator {

    // CDI injections won't work here because it's not a CDI bean.

    @Override
    public Object generate(Method method, Object... methodParams) {
        return new CompositeCacheKey(method.getName(), methodParams[0]); (1)
    }
}
1 Including the method name into the cache key is not expensive, unlike other methods from Method.

Both kinds of key generators can be used in a similar way:

package org.acme.cache;

import javax.enterprise.context.ApplicationScoped;

import org.acme.cache.ApplicationScopedKeyGen;
import org.acme.cache.NotABeanKeyGen;

import io.quarkus.cache.CacheKey;
import io.quarkus.cache.CacheInvalidate;
import io.quarkus.cache.CacheResult;

@ApplicationScoped
public class CachedService {

    @CacheResult(cacheName = "foo", keyGenerator = ApplicationScopedKeyGen.class) (1)
    public Object load(@CacheKey Object notUsedInKey, String keyElement) { (2)
        // Call expensive service here.
    }

    @CacheInvalidate(cacheName = "foo", keyGenerator = NotABeanKeyGen.class) (3)
    public void invalidate(Object keyElement) {
    }

    @CacheInvalidate(cacheName = "foo", keyGenerator = NotABeanKeyGen.class)
    @CacheInvalidate(cacheName = "bar")
    public void invalidate(Integer param0, @CacheKey BigDecimal param1) { (4)
    }
}
1 This key generator is a CDI bean.
2 The @CacheKey annotation will be ignored because a key generator is declared in the @CacheResult annotation.
3 This key generator is not a CDI bean.
4 The @CacheKey annotation will be ignored when the foo cache data is invalidated, but param1 will be the cache key when the bar cache data is invalidated.

Caching using the programmatic API

Quarkus also offers a programmatic API which can be used to store, retrieve or delete values from any cache declared using the annotations API. All operations from the programmatic API are non-blocking and rely on Mutiny under the hood.

Before accessing programmatically the cached data, you need to retrieve an io.quarkus.cache.Cache instance. The following sections will show you how to do that.

Injecting a Cache with the @CacheName annotation

io.quarkus.cache.CacheName can be used on a field, a constructor parameter or a method parameter to inject a Cache:

package org.acme.cache;

import javax.enterprise.context.ApplicationScoped;
import javax.inject.Inject;

import io.quarkus.cache.Cache;
import io.quarkus.cache.CacheName;
import io.smallrye.mutiny.Uni;

@ApplicationScoped
public class CachedExpensiveService {

    @Inject (1)
    @CacheName("my-cache")
    Cache cache;

    public Uni<String> getNonBlockingExpensiveValue(Object key) { (2)
        return cache.get(key, k -> { (3)
            /*
             * Put an expensive call here.
             * It will be executed only if the key is not already associated with a value in the cache.
             */
        });
    }

    public String getBlockingExpensiveValue(Object key) {
        return cache.get(key, k -> {
            // Put an expensive call here.
        }).await().indefinitely(); (4)
    }
}
1 This is optional.
2 This method returns the Uni<String> type which is non-blocking.
3 The k argument contains the cache key value.
4 If you don’t need the call to be non-blocking, this is how you can retrieve the cache value in a blocking way.

Retrieving a Cache from the CacheManager

Another way to retrieve a Cache instance consists in injecting the io.quarkus.cache.CacheManager first and then retrieving the desired Cache from its name:

package org.acme.cache;

import javax.enterprise.context.ApplicationScoped;
import javax.inject.Inject;

import io.quarkus.cache.Cache;
import io.quarkus.cache.CacheManager;

import java.util.Optional;

@ApplicationScoped
public class CacheClearer {

    @Inject
    CacheManager cacheManager;

    public void clearCache(String cacheName) {
        Optional<Cache> cache = cacheManager.getCache(cacheName);
        if (cache.isPresent()) {
            cache.get().invalidateAll().await().indefinitely();
        }
    }
}

Building a programmatic cache key

Before building a programmatic cache key, you need to know how cache keys are built by the annotations API when an annotated method is invoked. This is explained in the Cache keys building logic section of this guide.

Now, if you want to retrieve or delete, using the programmatic API, a cache value that was stored using the annotations API, you just need to make sure the same key is used with both APIs.

Retrieving all keys from a CaffeineCache

The cache keys from a specific CaffeineCache can be retrieved as an unmodifiable Set as shown below. If the cache entries are modified while an iteration over the set is in progress, the set will remain unchanged.

package org.acme.cache;

import javax.enterprise.context.ApplicationScoped;

import io.quarkus.cache.Cache;
import io.quarkus.cache.CacheName;
import io.quarkus.cache.CaffeineCache;

import java.util.Set;

@ApplicationScoped
public class CacheKeysService {

    @CacheName("my-cache")
    Cache cache;

    public Set<Object> getAllCacheKeys() {
        return cache.as(CaffeineCache.class).keySet();
    }
}

Configuring the underlying caching provider

This extension uses Caffeine as its underlying caching provider. Caffeine is a high performance, near optimal caching library.

Caffeine configuration properties

Each of the Caffeine caches backing up the Quarkus application data caching extension can be configured using the following properties in the application.properties file. By default, caches do not perform any type of eviction if not configured.

You need to replace cache-name in all the following properties with the real name of the cache you want to configure.

Configuration property fixed at build time - All other configuration properties are overridable at runtime

Configuration property

类型

默认

Minimum total size for the internal data structures. Providing a large enough estimate at construction time avoids the need for expensive resizing operations later, but setting this value unnecessarily high wastes memory.

Environment variable: QUARKUS_CACHE_CAFFEINE__CACHE_NAME__INITIAL_CAPACITY

int

Maximum number of entries the cache may contain. Note that the cache may evict an entry before this limit is exceeded or temporarily exceed the threshold while evicting. As the cache size grows close to the maximum, the cache evicts entries that are less likely to be used again. For example, the cache may evict an entry because it hasn’t been used recently or very often.

Environment variable: QUARKUS_CACHE_CAFFEINE__CACHE_NAME__MAXIMUM_SIZE

long

Specifies that each entry should be automatically removed from the cache once a fixed duration has elapsed after the entry’s creation, or the most recent replacement of its value.

Environment variable: QUARKUS_CACHE_CAFFEINE__CACHE_NAME__EXPIRE_AFTER_WRITE

Duration

Specifies that each entry should be automatically removed from the cache once a fixed duration has elapsed after the entry’s creation, the most recent replacement of its value, or its last read.

Environment variable: QUARKUS_CACHE_CAFFEINE__CACHE_NAME__EXPIRE_AFTER_ACCESS

Duration

Environment variable: QUARKUS_CACHE_CAFFEINE__CACHE_NAME__METRICS_ENABLED

boolean

false

About the Duration format

持续时间的格式使用标准的 java.time.Duration 格式您可以在 Duration#parse() javadoc 中了解更多信息。

您还可以提供以数字开头的持续时间值。 在这种情况下,如果该值仅包含一个数字,则转换器将该值视为秒。 否则,PT 会隐式添加到值的前面,以获得标准的 java.time.Duration 格式。

Here’s what your cache configuration could look like:

quarkus.cache.caffeine."foo".initial-capacity=10 (1)
quarkus.cache.caffeine."foo".maximum-size=20
quarkus.cache.caffeine."foo".expire-after-write=60S
quarkus.cache.caffeine."bar".maximum-size=1000 (2)
1 The foo cache is being configured.
2 The bar cache is being configured.

Enabling Micrometer metrics

Each cache declared using the annotations caching API can be monitored using Micrometer metrics.

The cache metrics collection will only work if your application depends on a quarkus-micrometer-registry-* extension. See the Micrometer metrics guide to learn how to use Micrometer in Quarkus.

The cache metrics collection is disabled by default. It can be enabled from the application.properties file:

quarkus.cache.caffeine."foo".metrics-enabled=true

Like all instrumentation methods, collecting metrics comes with a small overhead that can impact the application performances.

The collected metrics contain cache statistics such as:

  • the approximate current number of entries in the cache

  • the number of entries that were added to the cache

  • the number of times a cache lookup has been performed, including information about hits and misses

  • the number of evictions and the weight of the evicted entries

Here is an example of cache metrics available for an application that depends on the quarkus-micrometer-registry-prometheus extension:

# HELP cache_size The number of entries in this cache. This may be an approximation, depending on the type of cache.
# TYPE cache_size gauge
cache_size{cache="foo",} 8.0
# HELP cache_puts_total The number of entries added to the cache
# TYPE cache_puts_total counter
cache_puts_total{cache="foo",} 12.0
# HELP cache_gets_total The number of times cache lookup methods have returned a cached value.
# TYPE cache_gets_total counter
cache_gets_total{cache="foo",result="hit",} 53.0
cache_gets_total{cache="foo",result="miss",} 12.0
# HELP cache_evictions_total cache evictions
# TYPE cache_evictions_total counter
cache_evictions_total{cache="foo",} 4.0
# HELP cache_eviction_weight_total The sum of weights of evicted entries. This total does not include manual invalidations.
# TYPE cache_eviction_weight_total counter
cache_eviction_weight_total{cache="foo",} 540.0

Annotated beans examples

Implicit simple cache key

package org.acme.cache;

import javax.enterprise.context.ApplicationScoped;

import io.quarkus.cache.CacheInvalidate;
import io.quarkus.cache.CacheInvalidateAll;
import io.quarkus.cache.CacheResult;

@ApplicationScoped
public class CachedService {

    @CacheResult(cacheName = "foo")
    public Object load(Object key) { (1)
        // Call expensive service here.
    }

    @CacheInvalidate(cacheName = "foo")
    public void invalidate(Object key) { (1)
    }

    @CacheInvalidateAll(cacheName = "foo")
    public void invalidateAll() {
    }
}
1 The cache key is implicit since there’s no @CacheKey annotation.

Explicit composite cache key

package org.acme.cache;

import javax.enterprise.context.Dependent;

import io.quarkus.cache.CacheInvalidate;
import io.quarkus.cache.CacheInvalidateAll;
import io.quarkus.cache.CacheKey;
import io.quarkus.cache.CacheResult;

@Dependent
public class CachedService {

    @CacheResult(cacheName = "foo")
    public String load(@CacheKey Object keyElement1, @CacheKey Object keyElement2, Object notPartOfTheKey) { (1)
        // Call expensive service here.
    }

    @CacheInvalidate(cacheName = "foo")
    public void invalidate(@CacheKey Object keyElement1, @CacheKey Object keyElement2, Object notPartOfTheKey) { (1)
    }

    @CacheInvalidateAll(cacheName = "foo")
    public void invalidateAll() {
    }
}
1 The cache key is explicitly composed of two elements. The method signature also contains a third argument which is not part of the key.

Default cache key

package org.acme.cache;

import javax.enterprise.context.Dependent;

import io.quarkus.cache.CacheInvalidate;
import io.quarkus.cache.CacheInvalidateAll;
import io.quarkus.cache.CacheResult;

@Dependent
public class CachedService {

    @CacheResult(cacheName = "foo")
    public String load() { (1)
        // Call expensive service here.
    }

    @CacheInvalidate(cacheName = "foo")
    public void invalidate() { (1)
    }

    @CacheInvalidateAll(cacheName = "foo")
    public void invalidateAll() {
    }
}
1 A unique default cache key derived from the cache name is used because the method has no arguments.

Multiple annotations on a single method

package org.acme.cache;

import javax.inject.Singleton;

import io.quarkus.cache.CacheInvalidate;
import io.quarkus.cache.CacheInvalidateAll;
import io.quarkus.cache.CacheResult;

@Singleton
public class CachedService {

    @CacheInvalidate(cacheName = "foo")
    @CacheResult(cacheName = "foo")
    public String forceCacheEntryRefresh(Object key) { (1)
        // Call expensive service here.
    }

    @CacheInvalidateAll(cacheName = "foo")
    @CacheInvalidateAll(cacheName = "bar")
    public void multipleInvalidateAll(Object key) { (2)
    }
}
1 This method can be used to force a refresh of the cache entry corresponding to the given key.
2 This method will invalidate all entries from the foo and bar caches with a single call.

Clear all application caches

package org.acme.cache;

import javax.enterprise.context.ApplicationScoped;
import javax.inject.Inject;

import io.quarkus.cache.CacheManager;

@ApplicationScoped
public class CacheClearer {

    @Inject
    CacheManager cacheManager;

    public void clearAllCaches() {
        for (String cacheName : cacheManager.getCacheNames()) {
            cacheManager.getCache(cacheName).get().invalidateAll().await().indefinitely();
        }
    }
}

Negative caching and nulls

Sometimes one wants to cache the result of an (expensive) remote call. If the remote call fails, one may not want to cache the result or exception, but rather re-try the remote call on the next invocation.

A simple approach could be to catch the exception and return null, so that the caller can act accordingly:

Sample code
    public void caller(int val) {

        Integer result = callRemote(val); (1)
        if (result != null) {
            System.out.println("Result is " + result);
        else {
            System.out.println("Got an exception");
        }
    }

    @CacheResult(cacheName = "foo")
    public Integer callRemote(int val)  {

        try {
            Integer val = remoteWebServer.getResult(val); (2)
            return val;
        } catch (Exception e) {
            return null; (3)
        }
    }
1 Call the method to call the remote
2 Do the remote call and return its result
3 Return in case of exception

This approach has an unfortunate side effect: as we said before, Quarkus can also cache null values. Which means that the next call to callRemote() with the same parameter value will be answered out of the cache, returning null and no remote call will be done. This may be desired in some scenarios, but usually one wants to retry the remote call until it returns a result.

Let exceptions bubble up

To prevent the cache from caching (marker) results from a remote call, we need to let the exception bubble out of the called method and catch it at the caller side:

With Exception bubbling up
   public void caller(int val) {
       try {
           Integer result = callRemote(val);  (1)
           System.out.println("Result is " + result);
       } catch (Exception e) {
           System.out.println("Got an exception");
   }

   @CacheResult(cacheName = "foo")
   public Integer callRemote(int val) throws Exception { (2)

      Integer val = remoteWebServer.getResult(val);  (3)
      return val;

   }
1 Call the method to call the remote
2 Exceptions may bubble up
3 This can throw all kinds of remote exceptions

When the call to the remote throws an exception, the cache does not store the result, so that a subsequent call to callRemote() with the same parameter value will not be answered out of the cache. It will instead result in another attempt to call the remote.

Going native

The Cache extension supports building native executables.

However, to optimize runtime speed, Caffeine embarks many cache implementation classes that are selected depending on the cache configuration. We are not registering all of them for reflection (and the ones not registered are not included into the native executables) as registering all of them would be very costly.

We are registering the most common implementations but, depending on your cache configuration, you might encounter errors like:

2021-12-08 02:32:02,108 ERROR [io.qua.run.Application] (main) Failed to start application (with profile prod): java.lang.ClassNotFoundException: com.github.benmanes.caffeine.cache.PSAMS (1)
        at java.lang.Class.forName(DynamicHub.java:1433)
        at java.lang.Class.forName(DynamicHub.java:1408)
        at com.github.benmanes.caffeine.cache.NodeFactory.newFactory(NodeFactory.java:111)
        at com.github.benmanes.caffeine.cache.BoundedLocalCache.<init>(BoundedLocalCache.java:240)
        at com.github.benmanes.caffeine.cache.SS.<init>(SS.java:31)
        at com.github.benmanes.caffeine.cache.SSMS.<init>(SSMS.java:64)
        at com.github.benmanes.caffeine.cache.SSMSA.<init>(SSMSA.java:43)
1 PSAMS is one of the many cache implementation classes of Caffeine so this part may vary.

When you encounter this error, you can easily fix it by adding the following annotation to any of your application classes (or you can create a new class such as Reflections just to host this annotation if you prefer):

@RegisterForReflection(classNames = { "com.github.benmanes.caffeine.cache.PSAMS" }) (1)
1 It is an array, so you can register several cache implementations in one go if your configuration requires several of them.

This annotation will register the cache implementation classes for reflection and this will include the classes into the native executable.