Redis Cloud: The Official Managed Redis

What Redis Cloud is, in 2026

Redis Cloud is the cloud-managed offering operated by Redis Inc. (the company that maintains Redis itself). It runs on AWS, Google Cloud, and Azure as the underlying infrastructure, but the management plane, billing, and support are direct from Redis Inc. Customers choose a region, a deployment topology (single-region or active-active multi-region), a memory size, and a list of optional modules. Redis Inc. provisions, monitors, patches, and operates the cluster.

The proposition versus ElastiCache, Memorystore, or self-hosted Redis is access to the Redis modules. RediSearch provides full-text and vector search. RedisJSON provides native JSON storage with JSONPath queries. RedisTimeSeries handles time-series data. RedisBloom adds probabilistic data structures (Bloom filters, Top-K, Count-Min). RedisGraph (now deprecated in favour of native graph features) and RedisGears (server-side functions). These modules are first-party Redis Inc. products that are technically possible to run on ElastiCache or self-host (under varying license terms) but officially supported only on Redis Cloud and Redis Enterprise.

The other proposition is the active-active geo-replication via CRDB. Multiple Redis Cloud clusters in different regions accept writes simultaneously, with CRDT-based conflict resolution. The cluster topology is a logical "Active-Active Database" that spans regions, with the application connecting to any of them and seeing eventual convergence across all. ElastiCache global datastores offer a similar primitive but with one writer at a time; Redis Enterprise CRDB allows truly concurrent multi-region writes.

The three tiers and the pricing model

Essentials is the starter tier: a fixed-size database (from a free 30MB up through paid sizes), single shard, single region, but with all the modules included. Pricing is per-database-per-month, scaling roughly linearly with database size. Suitable for development, small production workloads, and stateful evaluation. The free 30MB tier is genuinely useful for experimentation with the modules; it is one of the better free tiers among managed databases.

Pro is the production tier: pay-as-you-go pricing based on throughput (operations per second) and dataset size, with the ability to scale up and down on demand. Pro includes high availability, advanced security (TLS, IP allowlisting, VPC peering), and the full module suite. It is broadly comparable to other managed Redis services on price for similar capability, with the module access being the main differentiator.

Annual is the enterprise tier: committed contracts, custom pricing, active-active CRDB across regions, Redis on Flash (NVMe-tiered storage for very large datasets), and dedicated support. This is the tier that competes most directly with Redis Enterprise on-premises licensing. Pricing requires sales contact and is typically in the tens-to-hundreds-of-thousands-of-dollars annual range for substantial deployments.

Redis modules: the actual selling point

The Redis modules are why teams pick Redis Cloud over equivalent managed offerings. RediSearch adds full-text search (BM25-ranked, with synonyms, stemming, and aggregations) and vector similarity search (HNSW indexes, cosine / inner product / L2 distance) to Redis. A typical use case is RAG (retrieval-augmented generation) for AI workloads where embeddings are stored as Redis vector fields and queried by similarity at sub-millisecond latency.

RedisJSON stores native JSON documents with JSONPath query support: JSON.SET, JSON.GET, JSON.ARRAPPEND, JSON.NUMINCRBY. The serialised binary format is more compact and faster to query than stringified JSON in plain Redis strings. RedisTimeSeries adds time-series operations (downsampling, aggregation, retention policies). RedisBloom provides Bloom filters, Cuckoo filters, Top-K, and Count-Min sketches as first-class data types for memory-efficient probabilistic membership and frequency queries.

The modules can be installed on self-hosted Redis as well, but the licensing has historically been a Redis-Source-Available-License variant that prohibits competing managed services. The May 2025 AGPL addition gave a third option for the modules' license. The pragmatic reality is that the modules work best on Redis Cloud or Redis Enterprise because Redis Inc. supports them there as a first-party offering. Self-hosting the modules is possible and legal under the new tri-license but requires more operational lift.

When Redis Cloud is the right answer

The clear cases for Redis Cloud are: you want RediSearch for vector search or full-text search and you do not want to operate it yourself; you want active-active geo-replication for a global multi-write application; you want the latest Redis 8.0 features (vector sets, hash field TTL) and your cloud provider has not added them yet; you want first-party support from Redis Inc. for production debugging. For any of these, Redis Cloud is the most direct path.

The cases against are mostly price-driven. For a generic Redis workload (caching, sessions, rate limiting, queue) without specific module requirements, Redis Cloud is typically 20-40% more expensive per GB-month than ElastiCache or Memorystore on equivalent SKUs. The premium is the cost of the module bundle and the active-active capability; if you do not need them, you are paying for unused features.

The decision-tree shortcut: do you need RediSearch / vector / active-active? If yes, Redis Cloud or Redis Enterprise on-premises are the only paths. If no, pick the cheapest managed Redis your cloud provider offers, which is usually Memorystore for Valkey or ElastiCache for Valkey. Memcached as a candidate does not apply here because Redis Cloud does not offer it and the workloads where you would choose Redis Cloud are exactly the workloads where Memcached's lack of features would rule it out.

FAQ

Related decisions