One can find various samples for key-value stores in the dedicated Spring Data example repo, at https://github.com/spring-projects/spring-data-examples/. See its documentation, the following blog entry for more information.

Spring Boot selects a recent version of the Spring Data modules for you. If you still want to upgrade to a newer version, set the spring-data-bom.version property to the train version and iteration you would like to use.

See Spring Boot’s documentation (search for "Spring Data Bom") for more details.

The current version of Spring Data modules require Spring Framework 6.0.7 or better. The modules might also work with an older bugfix version of that minor version. However, using the most recent version within that generation is highly recommended.

This part of the reference documentation explains the core functionality offered by Spring Data Redis. It explains Key-Value module concepts and semantics and the syntax for various stores namespaces. For an introduction to key-value stores, Spring, or Spring Data examples, see Learning NoSQL and Key Value Stores. This documentation refers only to Spring Data Redis Support and assumes the user is familiar with key-value storage and Spring concepts.

“Redis support” introduces the Redis module feature set.

“Redis Repositories” introduces the repository support for Redis.

This document is the reference guide for Spring Data Redis (SDR) Support.

Once you’ve decided to upgrade your application, you can find detailed information regarding specific features in the rest of the document. You can find migration guides specific to major version migrations at the end of this document.

Spring Data’s documentation is specific to that version, so any information that you find in here will contain the most up-to-date changes that are in that version.

An easy way to setting up a working environment is to create a Spring-based project in STS.

First, you need to set up a running Redis server.

To create a Spring project in STS:

The repository is also browseable here.

Spring Redis requires Redis 2.6 or above and Spring Data Redis integrates with Lettuce and Jedis, two popular open-source Java libraries for Redis.

The Redis support provides several components. For most tasks, the high-level abstractions and support services are the best choice. Note that, at any point, you can move between layers. For example, you can get a low-level connection (or even the native library) to communicate directly with Redis.

One of the first tasks when using Redis and Spring is to connect to the store through the IoC container. To do that, a Java connector (or binding) is required. No matter the library you choose, you need to use only one set of Spring Data Redis APIs (which behaves consistently across all connectors): the org.springframework.data.redis.connection package and its RedisConnection and RedisConnectionFactory interfaces for working with and retrieving active connections to Redis.

For dealing with high-availability Redis, Spring Data Redis has support for Redis Sentinel, using RedisSentinelConfiguration, as shown in the following example:

Sometimes, direct interaction with one of the Sentinels is required. Using RedisConnectionFactory.getSentinelConnection() or RedisConnection.getSentinelCommands() gives you access to the first active Sentinel configured.

Most users are likely to use RedisTemplate and its corresponding package, org.springframework.data.redis.core. The template is, in fact, the central class of the Redis module, due to its rich feature set. The template offers a high-level abstraction for Redis interactions. While RedisConnection offers low-level methods that accept and return binary values (byte arrays), the template takes care of serialization and connection management, freeing the user from dealing with such details.

Moreover, the template provides operations views (following the grouping from the Redis command reference) that offer rich, generified interfaces for working against a certain type or certain key (through the KeyBound interfaces) as described in the following table:

Once configured, the template is thread-safe and can be reused across multiple instances.

RedisTemplate uses a Java-based serializer for most of its operations. This means that any object written or read by the template is serialized and deserialized through Java. You can change the serialization mechanism on the template, and the Redis module offers several implementations, which are available in the org.springframework.data.redis.serializer package. See Serializers for more information. You can also set any of the serializers to null and use RedisTemplate with raw byte arrays by setting the enableDefaultSerializer property to false. Note that the template requires all keys to be non-null. However, values can be null as long as the underlying serializer accepts them. Read the Javadoc of each serializer for more information.

For cases where you need a certain template view, declare the view as a dependency and inject the template. The container automatically performs the conversion, eliminating the opsFor[X] calls, as shown in the following example:

Since it is quite common for the keys and values stored in Redis to be java.lang.String, the Redis modules provides two extensions to RedisConnection and RedisTemplate, respectively the StringRedisConnection (and its DefaultStringRedisConnection implementation) and StringRedisTemplate as a convenient one-stop solution for intensive String operations. In addition to being bound to String keys, the template and the connection use the StringRedisSerializer underneath, which means the stored keys and values are human-readable (assuming the same encoding is used both in Redis and your code). The following listings show an example:

As with the other Spring templates, RedisTemplate and StringRedisTemplate let you talk directly to Redis through the RedisCallback interface. This feature gives complete control to you, as it talks directly to the RedisConnection. Note that the callback receives an instance of StringRedisConnection when a StringRedisTemplate is used. The following example shows how to use the RedisCallback interface:

From the framework perspective, the data stored in Redis is only bytes. While Redis itself supports various types, for the most part, these refer to the way the data is stored rather than what it represents. It is up to the user to decide whether the information gets translated into strings or any other objects.

In Spring Data, the conversion between the user (custom) types and raw data (and vice-versa) is handled Redis in the org.springframework.data.redis.serializer package.

This package contains two types of serializers that, as the name implies, take care of the serialization process:

The main difference between these variants is that RedisSerializer primarily serializes to byte[] while readers and writers use ByteBuffer.

Multiple implementations are available (including two that have been already mentioned in this documentation):

However one can use OxmSerializer for Object/XML mapping through Spring OXM support or Jackson2JsonRedisSerializer or GenericJackson2JsonRedisSerializer for storing data in JSON format.

Do note that the storage format is not limited only to values. It can be used for keys, values, or hashes without any restrictions.

Data can be stored by using various data structures within Redis. Jackson2JsonRedisSerializer can convert objects in JSON format. Ideally, JSON can be stored as a value by using plain keys. You can achieve a more sophisticated mapping of structured objects by using Redis hashes. Spring Data Redis offers various strategies for mapping data to hashes (depending on the use case):

Spring Data provides dedicated messaging integration for Redis, similar in functionality and naming to the JMS integration in Spring Framework.

Redis messaging can be roughly divided into two areas of functionality:

This is an example of the pattern often called Publish/Subscribe (Pub/Sub for short). The RedisTemplate class is used for message production. For asynchronous reception similar to Java EE’s message-driven bean style, Spring Data provides a dedicated message listener container that is used to create Message-Driven POJOs (MDPs) and, for synchronous reception, the RedisConnection contract.

The org.springframework.data.redis.connection and org.springframework.data.redis.listener packages provide the core functionality for Redis messaging.

Redis Streams model a log data structure in an abstract approach. Typically, logs are append-only data structures and are consumed from the beginning on, at a random position, or by streaming new messages.

Redis Streams can be roughly divided into two areas of functionality:

Although this pattern has similarities to Pub/Sub, the main difference lies in the persistence of messages and how they are consumed.

While Pub/Sub relies on the broadcasting of transient messages (i.e. if you don’t listen, you miss a message), Redis Stream use a persistent, append-only data type that retains messages until the stream is trimmed. Another difference in consumption is that Pub/Sub registers a server-side subscription. Redis pushes arriving messages to the client while Redis Streams require active polling.

The org.springframework.data.redis.connection and org.springframework.data.redis.stream packages provide the core functionality for Redis Streams.

Redis provides support for transactions through the multi, exec, and discard commands. These operations are available on RedisTemplate. However, RedisTemplate is not guaranteed to run all the operations in the transaction with the same connection.

Spring Data Redis provides the SessionCallback interface for use when multiple operations need to be performed with the same connection, such as when using Redis transactions.The following example uses the multi method:

RedisTemplate uses its value, hash key, and hash value serializers to deserialize all results of exec before returning. There is an additional exec method that lets you pass a custom serializer for transaction results.

Redis provides support for pipelining, which involves sending multiple commands to the server without waiting for the replies and then reading the replies in a single step. Pipelining can improve performance when you need to send several commands in a row, such as adding many elements to the same List.

Spring Data Redis provides several RedisTemplate methods for running commands in a pipeline. If you do not care about the results of the pipelined operations, you can use the standard execute method, passing true for the pipeline argument. The executePipelined methods run the provided RedisCallback or SessionCallback in a pipeline and return the results, as shown in the following example:

The preceding example runs a bulk right pop of items from a queue in a pipeline. The results List contains all of the popped items. RedisTemplate uses its value, hash key, and hash value serializers to deserialize all results before returning, so the returned items in the preceding example are Strings. There are additional executePipelined methods that let you pass a custom serializer for pipelined results.

Note that the value returned from the RedisCallback is required to be null, as this value is discarded in favor of returning the results of the pipelined commands.

Redis versions 2.6 and higher provide support for running Lua scripts through the eval and evalsha commands. Spring Data Redis provides a high-level abstraction for running scripts that handles serialization and automatically uses the Redis script cache.

Scripts can be run by calling the execute methods of RedisTemplate and ReactiveRedisTemplate. Both use a configurable ScriptExecutor (or ReactiveScriptExecutor) to run the provided script. By default, the ScriptExecutor (or ReactiveScriptExecutor) takes care of serializing the provided keys and arguments and deserializing the script result. This is done through the key and value serializers of the template. There is an additional overload that lets you pass custom serializers for the script arguments and the result.

The default ScriptExecutor optimizes performance by retrieving the SHA1 of the script and attempting first to run evalsha, falling back to eval if the script is not yet present in the Redis script cache.

The following example runs a common “check-and-set” scenario by using a Lua script. This is an ideal use case for a Redis script, as it requires that running a set of commands atomically, and the behavior of one command is influenced by the result of another.

The preceding code configures a RedisScript pointing to a file called checkandset.lua, which is expected to return a boolean value. The script resultType should be one of Long, Boolean, List, or a deserialized value type. It can also be null if the script returns a throw-away status (specifically, OK).

The checkAndSet method above then runs the scripts. Scripts can be run within a SessionCallback as part of a transaction or pipeline. See “Redis Transactions” and “Pipelining” for more information.

The scripting support provided by Spring Data Redis also lets you schedule Redis scripts for periodic running by using the Spring Task and Scheduler abstractions. See the Spring Framework documentation for more details.

Spring Redis provides an implementation for the Spring cache abstraction through the org.springframework.data.redis.cache package. To use Redis as a backing implementation, add RedisCacheManager to your configuration, as follows:

RedisCacheManager behavior can be configured with RedisCacheManagerBuilder, letting you set the default RedisCacheConfiguration, transaction behavior, and predefined caches.

As shown in the preceding example, RedisCacheManager allows definition of configurations on a per-cache basis.

The behavior of RedisCache created with RedisCacheManager is defined with RedisCacheConfiguration. The configuration lets you set key expiration times, prefixes, and RedisSerializer implementations for converting to and from the binary storage format, as shown in the following example:

RedisCacheManager defaults to a lock-free RedisCacheWriter for reading and writing binary values. Lock-free caching improves throughput. The lack of entry locking can lead to overlapping, non-atomic commands for the putIfAbsent and clean methods, as those require multiple commands to be sent to Redis. The locking counterpart prevents command overlap by setting an explicit lock key and checking against presence of this key, which leads to additional requests and potential command wait times.

Locking applies on the cache level, not per cache entry.

It is possible to opt in to the locking behavior as follows:

By default, any key for a cache entry gets prefixed with the actual cache name followed by two colons. This behavior can be changed to a static as well as a computed prefix.

The following example shows how to set a static prefix:

The cache implementation defaults to use KEYS and DEL to clear the cache. KEYS can cause performance issues with large keyspaces. Therefore, the default RedisCacheWriter can be created with a BatchStrategy to switch to a SCAN-based batch strategy. The SCAN strategy requires a batch size to avoid excessive Redis command roundtrips:

The following table lists the default settings for RedisCacheManager:

The following table lists the default settings for RedisCacheConfiguration:

Package org.springframework.data.redis.support offers various reusable components that rely on Redis as a backing store. Currently, the package contains various JDK-based interface implementations on top of Redis, such as atomic counters and JDK Collections.

The atomic counters make it easy to wrap Redis key incrementation while the collections allow easy management of Redis keys with minimal storage exposure or API leakage. In particular, the RedisSet and RedisZSet interfaces offer easy access to the set operations supported by Redis, such as intersection and union. RedisList implements the List, Queue, and Deque contracts (and their equivalent blocking siblings) on top of Redis, exposing the storage as a FIFO (First-In-First-Out), LIFO (Last-In-First-Out) or capped collection with minimal configuration. The following example shows the configuration for a bean that uses a RedisList:

The following example shows a Java configuration example for a Deque:

As shown in the preceding example, the consuming code is decoupled from the actual storage implementation. In fact, there is no indication that Redis is used underneath. This makes moving from development to production environments transparent and highly increases testability (the Redis implementation can be replaced with an in-memory one).

Getting insights from an application component about its operations, timing and relation to application code is crucial to understand latency. Spring Data Redis ships with a Micrometer integration through the Lettuce driver to collect observations during Redis interaction. Once the integration is set up, Micrometer will create meters and spans (for distributed tracing) for each Redis command.

To enable the integration, apply the following configuration to LettuceClientConfiguration:

Spring Data Redis currently integrates with Lettuce as the only reactive Java connector. Project Reactor is used as reactive composition library.

One of the first tasks when using Redis and Spring is to connect to the store through the IoC container. To do that, a Java connector (or binding) is required. No matter the library you choose, you must use the org.springframework.data.redis.connection package and its ReactiveRedisConnection and ReactiveRedisConnectionFactory interfaces to work with and retrieve active connections to Redis.

Most users are likely to use ReactiveRedisTemplate and its corresponding package, org.springframework.data.redis.core. Due to its rich feature set, the template is, in fact, the central class of the Redis module. The template offers a high-level abstraction for Redis interactions. While ReactiveRedisConnection offers low-level methods that accept and return binary values (ByteBuffer), the template takes care of serialization and connection management, freeing you from dealing with such details.

Moreover, the template provides operation views (following the grouping from Redis command reference) that offer rich, generified interfaces for working against a certain type as described in the following table:

Once configured, the template is thread-safe and can be reused across multiple instances.

ReactiveRedisTemplate uses a Java-based serializer for most of its operations. This means that any object written or read by the template is serialized or deserialized through RedisElementWriter or RedisElementReader. The serialization context is passed to the template upon construction, and the Redis module offers several implementations available in the org.springframework.data.redis.serializer package. See Serializers for more information.

The following example shows a ReactiveRedisTemplate being used to return a Mono:

Since it is quite common for keys and values stored in Redis to be a java.lang.String, the Redis module provides a String-based extension to ReactiveRedisTemplate: ReactiveStringRedisTemplate. It is a convenient one-stop solution for intensive String operations. In addition to being bound to String keys, the template uses the String-based RedisSerializationContext, which means the stored keys and values are human readable (assuming the same encoding is used in both Redis and your code). The following example shows ReactiveStringRedisTemplate in use:

Spring Data provides dedicated messaging integration for Redis, very similar in functionality and naming to the JMS integration in Spring Framework; in fact, users familiar with the JMS support in Spring should feel right at home.

Redis messaging can be roughly divided into two areas of functionality, namely the production or publication and consumption or subscription of messages, hence the shortcut pubsub (Publish/Subscribe). The ReactiveRedisTemplate class is used for message production. For asynchronous reception, Spring Data provides a dedicated message listener container that is used consume a stream of messages. For the purpose of just subscribing ReactiveRedisTemplate offers stripped down alternatives to utilizing a listener container.

The package org.springframework.data.redis.connection and org.springframework.data.redis.listener provide the core functionality for using Redis messaging.

You can run Redis scripts with the reactive infrastructure by using the ReactiveScriptExecutor, which is best accessed through ReactiveRedisTemplate.

See to the scripting section for more details on scripting commands.

Cluster support is based on the same building blocks as non-clustered communication. RedisClusterConnection, an extension to RedisConnection, handles the communication with the Redis Cluster and translates errors into the Spring DAO exception hierarchy. RedisClusterConnection instances are created with the RedisConnectionFactory, which has to be set up with the associated RedisClusterConfiguration, as shown in the following example:

As mentioned earlier, Redis Cluster behaves differently from single-node Redis or even a Sentinel-monitored master-replica environment. This is because the automatic sharding maps a key to one of 16384 slots, which are distributed across the nodes. Therefore, commands that involve more than one key must assert all keys map to the exact same slot to avoid cross-slot errors. A single cluster node serves only a dedicated set of keys. Commands issued against one particular server return results only for those keys served by that server. As a simple example, consider the KEYS command. When issued to a server in a cluster environment, it returns only the keys served by the node the request is sent to and not necessarily all keys within the cluster. So, to get all keys in a cluster environment, you must read the keys from all the known master nodes.

While redirects for specific keys to the corresponding slot-serving node are handled by the driver libraries, higher-level functions, such as collecting information across nodes or sending commands to all nodes in the cluster, are covered by RedisClusterConnection. Picking up the keys example from earlier, this means that the keys(pattern) method picks up every master node in the cluster and simultaneously runs the KEYS command on every master node while picking up the results and returning the cumulated set of keys. To just request the keys of a single node RedisClusterConnection provides overloads for those methods (for example, keys(node, pattern)).

A RedisClusterNode can be obtained from RedisClusterConnection.clusterGetNodes or it can be constructed by using either the host and the port or the node Id.

The following example shows a set of commands being run across the cluster:

When all keys map to the same slot, the native driver library automatically serves cross-slot requests, such as MGET. However, once this is not the case, RedisClusterConnection runs multiple parallel GET commands against the slot-serving nodes and again returns an accumulated result. This is less performant than the single-slot approach and, therefore, should be used with care. If in doubt, consider pinning keys to the same slot by providing a prefix in curly brackets, such as {my-prefix}.thing1 and {my-prefix}.thing2, which will both map to the same slot number. The following example shows cross-slot request handling:

See the Working with Objects through RedisTemplate section for information about the general purpose, configuration, and usage of RedisTemplate.

RedisTemplate provides access to cluster-specific operations through the ClusterOperations interface, which can be obtained from RedisTemplate.opsForCluster(). This lets you explicitly run commands on a single node within the cluster while retaining the serialization and deserialization features configured for the template. It also provides administrative commands (such as CLUSTER MEET) or more high-level operations (for example, resharding).

The following example shows how to access RedisClusterConnection with RedisTemplate:

Spring Data Redis lets you easily implement domain entities, as shown in the following example:

We have a pretty simple domain object here. Note that it has a @RedisHash annotation on its type and a property named id that is annotated with org.springframework.data.annotation.Id. Those two items are responsible for creating the actual key used to persist the hash.

To now actually have a component responsible for storage and retrieval, we need to define a repository interface, as shown in the following example:

As our repository extends CrudRepository, it provides basic CRUD and finder operations. The thing we need in between to glue things together is the corresponding Spring configuration, shown in the following example:

Given the preceding setup, we can inject PersonRepository into our components, as shown in the following example:

This section covers the fundamentals of Spring Data object mapping, object creation, field and property access, mutability and immutability. Note, that this section only applies to Spring Data modules that do not use the object mapping of the underlying data store (like JPA). Also be sure to consult the store-specific sections for store-specific object mapping, like indexes, customizing column or field names or the like.

Core responsibility of the Spring Data object mapping is to create instances of domain objects and map the store-native data structures onto those. This means we need two fundamental steps:

The Redis Repository support persists Objects to Hashes. This requires an Object-to-Hash conversion which is done by a RedisConverter. The default implementation uses Converter for mapping property values to and from Redis native byte[].

Given the Person type from the previous sections, the default mapping looks like the following:

The following table describes the default mapping rules:

Mapping behavior can be customized by registering the corresponding Converter in RedisCustomConversions. Those converters can take care of converting from and to a single byte[] as well as Map<String,byte[]>. The first one is suitable for (for example) converting a complex type to (for example) a binary JSON representation that still uses the default mappings hash structure. The second option offers full control over the resulting hash.

The following example shows two sample byte array converters:

Using the preceding byte array Converter produces output similar to the following:

The following example shows two examples of Map converters:

Using the preceding Map Converter produces output similar to the following:

Keyspaces define prefixes used to create the actual key for the Redis Hash. By default, the prefix is set to getClass().getName(). You can alter this default by setting @RedisHash on the aggregate root level or by setting up a programmatic configuration. However, the annotated keyspace supersedes any other configuration.

The following example shows how to set the keyspace configuration with the @EnableRedisRepositories annotation:

The following example shows how to programmatically set the keyspace:

Secondary indexes are used to enable lookup operations based on native Redis structures. Values are written to the according indexes on every save and are removed when objects are deleted or expire.

Objects stored in Redis may be valid only for a certain amount of time. This is especially useful for persisting short-lived objects in Redis without having to remove them manually when they reach their end of life. The expiration time in seconds can be set with @RedisHash(timeToLive=…​) as well as by using KeyspaceSettings (see Keyspaces).

More flexible expiration times can be set by using the @TimeToLive annotation on either a numeric property or a method. However, do not apply @TimeToLive on both a method and a property within the same class. The following example shows the @TimeToLive annotation on a property and on a method:

The repository implementation ensures subscription to Redis keyspace notifications via RedisMessageListenerContainer.

When the expiration is set to a positive value, the corresponding EXPIRE command is run. In addition to persisting the original, a phantom copy is persisted in Redis and set to expire five minutes after the original one. This is done to enable the Repository support to publish RedisKeyExpiredEvent, holding the expired value in Spring’s ApplicationEventPublisher whenever a key expires, even though the original values have already been removed. Expiry events are received on all connected applications that use Spring Data Redis repositories.

By default, the key expiry listener is disabled when initializing the application. The startup mode can be adjusted in @EnableRedisRepositories or RedisKeyValueAdapter to start the listener with the application or upon the first insert of an entity with a TTL. See EnableKeyspaceEvents for possible values.

The RedisKeyExpiredEvent holds a copy of the expired domain object as well as the key.

Marking properties with @Reference allows storing a simple key reference instead of copying values into the hash itself. On loading from Redis, references are resolved automatically and mapped back into the object, as shown in the following example:

In some cases, you need not load and rewrite the entire entity just to set a new value within it. A session timestamp for the last active time might be such a scenario where you want to alter one property. PartialUpdate lets you define set and delete actions on existing objects while taking care of updating potential expiration times of both the entity itself and index structures. The following example shows a partial update:

Query methods allow automatic derivation of simple finder queries from the method name, as shown in the following example:

Using derived query methods might not always be sufficient to model the queries to run. RedisCallback offers more control over the actual matching of index structures or even custom indexes. To do so, provide a RedisCallback that returns a single or Iterable set of id values, as shown in the following example:

The following table provides an overview of the keywords supported for Redis and what a method containing that keyword essentially translates to:

You can use the Redis repository support in a clustered Redis environment. See the “Redis Cluster” section for ConnectionFactory configuration details. Still, some additional configuration must be done, because the default key distribution spreads entities and secondary indexes through out the whole cluster and its slots.

The following table shows the details of data on a cluster (based on previous examples):

Instances of the repository interfaces are usually created by a container, for which Spring is the most natural choice when working with Spring Data. Spring offers sophisticated for creating bean instances. Spring Data Redis ships with a custom CDI extension that lets you use the repository abstraction in CDI environments. The extension is part of the JAR, so, to activate it, drop the Spring Data Redis JAR into your classpath.

You can then set up the infrastructure by implementing a CDI Producer for the RedisConnectionFactory and RedisOperations, as shown in the following example:

The necessary setup can vary, depending on your JavaEE environment.

The Spring Data Redis CDI extension picks up all available repositories as CDI beans and creates a proxy for a Spring Data repository whenever a bean of a repository type is requested by the container. Thus, obtaining an instance of a Spring Data repository is a matter of declaring an @Injected property, as shown in the following example:

A Redis Repository requires RedisKeyValueAdapter and RedisKeyValueTemplate instances. These beans are created and managed by the Spring Data CDI extension if no provided beans are found. You can, however, supply your own beans to configure the specific properties of RedisKeyValueAdapter and RedisKeyValueTemplate.

Redis as a store itself offers a very narrow low-level API leaving higher level functions, such as secondary indexes and query operations, up to the user.

This section provides a more detailed view of commands issued by the repository abstraction for a better understanding of potential performance implications.

Consider the following entity class as the starting point for all operations: