Redis Cloud

Managed Redis from the original creators

Redis Tutorial for Beginners #2 - Making a Redis Database (Redis Cloud)

3.6(0 reviews)freemiumFounded 2011United States

free tiermanaged optionbackups includedeu datacenterus datacenterapac datacenterapi access

Integrates with

AWS Terraform Redis

Redis Cloud is the official managed Redis service from Redis Ltd, the creators of Redis. Offers Redis Stack with JSON, search, and time series. Multi-cloud on AWS, GCP, Azure.

What Redis Cloud actually is

Redis Cloud is the managed Redis service from Redis Ltd. — the company behind Redis itself. The lineage matters here because it is genuinely unusual: Redis was written by Salvatore Sanfilippo in 2009, and the company that commercializes it was founded in 2011 (originally as Garantia Data, renamed Redis Labs in 2014, then Redis in 2021), with Sanfilippo joining to lead open-source development in 2015. When you buy Redis Cloud, you are buying from the stewards of the upstream project, and the managed service runs the same engine — not a reimplementation.

The product covers three deployment shapes: Redis Cloud Essentials (shared infrastructure, fixed-size plans), Redis Cloud Pro (dedicated infrastructure, unlimited RAM, multiple databases per subscription, Active-Active multi-region), and Redis Enterprise Software for self-managed or hybrid deployments. All of it runs on AWS, Google Cloud and Microsoft Azure — with Pro on Azure delivered through Azure Managed Redis. Underneath sits Redis 8 (May 2025), which collapsed the old Redis Stack module line into the core: JSON documents, the Redis Query Engine for search and vector workloads, time series and five probabilistic data structures ship in every binary distribution now. The name change with that release — Redis Community Edition became Redis Open Source — is also the licensing story: after the 2024 move to source-available terms, Redis 8 is offered under RSALv2, SSPLv1 or AGPLv3, your choice. Our category is /category/database-hosting.

Pricing: what you really pay

The published plan structure is clean, and we like that the per-hour math is shown on the pricing page. Free is $0.00 /hour: a 30 MB single database on shared infrastructure with best-effort SLA and community support. Essentials starts at $0.007 /hour, totaling $5/month for the smallest plan — shared deployment, 250 MB to 12 GB (the marketing page elsewhere says up to 100 GB; treat the plan table as authoritative and ask sales if you need the upper sizes), single database, SAML SSO, RBAC, encryption in transit and at rest, and up to 99.99% uptime. Pro starts at $0.014 /hour with a $200/month minimum and "first $200 free" framing: dedicated deployment, 6 GB RAM and up, unlimited RAM, multiple databases, Active-Active multi-region, auto-tiering, private connectivity (VPC peering, PrivateLink, Private Service Connect), and up to 99.999% uptime.

What does that mean in practice? A competitor's published cost analysis (Upstash, August 2026) works three examples from the current rate cards: a small 1 GB app at roughly $18/month on Essentials, a 10 GB production database with high availability at $549/month on Pro (their AWS Frankfurt example priced at $0.752/hour), and a multi-region Active-Active setup billing on the order of four times the data size. We have not verified those figures against an invoice, but the direction matches the published structure: the entry price is genuinely low and the production-grade configurations compound quickly. Data transfer is included in-plan (limited) on Essentials and Pro on-demand; Pro annual is usage-based for egress.

The fine print worth reading before you commit: on Essentials, database size includes replication, so a 1 GB plan with high availability yields about 512 MB of usable data. Throughput allowances assume small (roughly 1 KiB) requests — bigger values burn the budget faster. Per-plan limits on connections, bandwidth and ops are tiered (per the December 2025 tables: the 250 MB plan allows 256 connections and 1,000 ops/sec, up to 10,000 connections and 16,000 ops/sec at 12 GB). The Free plan is deliberately cramped: 30 MB, 100 operations per second, 30 concurrent connections, 5 GB of monthly bandwidth, one database, no replication and no backups. It is for learning, and Redis says so.

Performance & infrastructure

Redis does not need defending on latency — the engine's whole reason for existing is sub-millisecond data structures in memory. The managed service's job is to make that engine survive failure, and the architecture does the expected things: replication and automatic failover on paid plans, multi-AZ options, dedicated (Pro) versus shared (Essentials) placement, and for Pro an Active-Active mode with conflict-free multi-region writes. Redis's own historical claims about recovering from failures in seconds without data loss versus competitors losing data are vendor benchmarks (2015-era, Altoros study) — dated, but the failover machinery has been the product's differentiator since.

Region coverage is broad: databases deploy across AWS, Google Cloud and Azure regions on multiple continents (the supported-regions documentation organizes them per provider, including Europe, the Americas, Middle East/Africa and Asia-Pacific tabs; one AWS Middle East region was flagged temporarily unavailable at the time of writing). Pro on AWS supports VPC Peering, Transit Gateway and PrivateLink; Pro on GCP supports VPC Peering and Private Service Connect. Azure coverage for Essentials spans multiple regions.

Backups are a genuine inclusion on paid plans, and the specifics are documented: paid Essentials plans get scheduled backups every 24 hours plus on-demand backups and exports; Pro lets you set intervals of 24, 12, 6, 4, 2 or 1 hour with a chosen backup window and retention. Destinations include S3, Google Cloud Storage, Azure storage and FTP. Free plans cannot back up or export through the console at all — one more reason not to treat the free tier as production.

Tooling is strong. RedisInsight (the GUI that came from the 2019 RDBTools acquisition) is free and capable; a Terraform provider covers subscription provisioning; marketplaces on AWS, GCP and Azure let you burn committed cloud spend. And with Redis 8, the database is no longer just a cache: JSON documents, secondary-index search, vector sets (preview) and time series land in the same binary, which changes what belongs in Redis versus a separate store.

Where it falls short

Cost at production scale is the recurring complaint, and it is legitimate. In-memory data is expensive data; a 10 GB HA database at Pro's $200/month floor plus usage is real money versus running Redis yourself on a $20 VPS, and Active-Active multi-region multiplies storage costs. The unit prices are transparent; the architecture choices are what surprise people.

Second, the shared-tier ceilings. Essentials caps throughput, connections and bandwidth per plan size, and the RAM-includes-replication detail halves usable data at HA. Growing past 12 GB (or the advertised 100 GB — the conflicting published ceilings are a transparency wart) means moving to Pro and its $200/month minimum.

Third, licensing history. The 2024 switch from BSD to source-available terms was a real break with the project's earlier promise, and it matters for anyone embedding Redis or building services on it; the AGPLv3 option added with Redis 8 repaired some of that for operators who need an OSI license, but enterprises should read the license matrix (redis.io/legal/licenses) before assuming anything. For the managed service itself, lock-in risk is modest — it speaks the Redis protocol — but Pro's proprietary clustering and Active-Active features are not portable to a bare self-hosted Redis.

Finally, the free tier is among the weakest in this category (30 MB, no backups, no failover), and support tiers are sharply stratified: community support and best-effort SLA on Free, basic support on Essentials, and the fuller commitments only on Pro and Enterprise.

Who should pick this

We'd pick Redis Cloud when the cache or data store is business-critical and you want it operated by the people who write the engine: session stores, real-time leaderboards, vector search alongside existing Redis data, rate limiting, queues, and Active-Active setups serving multiple regions with writes in each. Teams standardized on AWS/GCP/Azure committed spend can buy through marketplaces, and the plan structure makes sense from a small Essentials database up to dedicated fleets.

We'd look elsewhere when: you are price-sensitive at moderate scale (self-hosted Redis or Valkey-class alternatives on /category/vps-providers hardware cost a fraction), you need a free tier for anything beyond learning, or you want one database spanning relational workloads — Redis is a specialized engine, not a general-purpose database. Our /benchmarks coverage of the underlying compute market is the right place to price the self-hosted alternative honestly.

Key takeaways

Managed Redis from Redis Ltd., the upstream stewards; Redis 8 (May 2025) folds JSON, search/query, time series and probabilistic structures into the core engine.
Free plan is real but tiny: 30 MB, 100 ops/sec, no replication, no backups — learning use only.
Essentials from $5/month (250 MB shared, up to 99.99% uptime) is the honest entry; Pro starts at a $200/month minimum (dedicated, Active-Active, up to 99.999%).
Paid plans include scheduled backups (daily on Essentials, down to hourly on Pro) to S3/GCS/Azure/FTP; Free has none.
Runs on AWS, Google Cloud and Azure across multiple continents, with private connectivity options on Pro.
Production costs compound — a 10 GB HA database lands in the several-hundred-dollars class; our overall view lands at 3.6/5.

Sources

Pros

Managed by the stewards of upstream Redis, running the same engine (Redis 8) rather than a fork
Real free tier for learning and a low $5/month Essentials entry with published per-hour math
Paid plans include scheduled backups (daily on Essentials, hourly-capable on Pro) to your own S3/GCS/Azure/FTP
Broad footprint: AWS, Google Cloud and Azure regions across multiple continents, private connectivity on Pro
Redis 8 turns the cache into a document, search, vector and time-series engine in one binary

Cons

Costs compound fast at production scale — in-memory data plus HA plus multi-region is expensive
Free plan is tiny (30 MB, 100 ops/sec, no replication, no backups) and Essentials caps throughput per tier
Essentials database size includes replication — HA halves usable capacity
Published storage ceilings conflict between pages (12 GB vs 100 GB for Essentials)
Licensing history (2024 source-available switch) requires reading the license matrix for embedded use

Best For

Teams running business-critical caching, session, real-time or vector workloads who want the Redis engine operated by its creators, with backups and SLAs included on paid plans.