Which plan to choose for Cache and Queues
Cache and Queues are managed services; Cache and Queues with Valkey are compatible with Redis clients and commands. For both, the plan's memory is the main resource: in Cache, it defines how much data is kept; in Queues, how many messages can wait for processing without putting pressure on the service.
Cache
Cache keeps data in memory to speed up the application: query results, sessions, external API responses and counters. The data can be discarded and recomputed, so the focus is speed, not durability.
| Plan | vCPU | RAM | Focus | Typical examples |
|---|---|---|---|---|
cache-free | 0.25 | 256 MiB | Get started at no cost | Prototype cache, sessions for a study project |
cache-starter | 0.5 | 512 MiB | Small projects and development environments | Login sessions, page cache for a small site |
cache-basic | 1 | 1 GiB | Growing applications | API query cache, per-user rate limiting |
cache-premium | 2 | 2 GiB | More data in memory, with metrics | Cache shared by several applications, real-time rankings and counters |
cache-enterprise | 8 | 8 GiB | Large volumes in memory | Platforms with many users and lots of hot data |
How to choose: estimate the size of what you want to keep and leave headroom. If the cache is always full and evicts keys that would still be useful, move to a larger plan in the Plan card of the service settings (hourly billing). The change restarts the cache, which comes back empty; cache-free does not change plan. See Change the plan.
Queues
Queues receive tasks and messages for asynchronous processing by workers, with persistent storage. You choose the engine when you create the queue: Valkey (simple and fast, with Streams and Pub/Sub) or RabbitMQ® (exchanges, routing keys, acknowledgments, and dead-letter). Both use the same plans and prices. See Queues: choose an engine, Queues with Valkey, and Queues with RabbitMQ.
| Plan | vCPU | RAM | Storage | Focus | Typical examples |
|---|---|---|---|---|---|
queue-free | 0.25 | 256 MiB | 1 GB fixed | Get started at no cost (Valkey only) | Prototype queue, learning Streams and Pub/Sub |
queue-starter | 0.5 | 512 MiB | 5 GB, expandable | First production queues | Sending emails, processing webhooks |
queue-basic | 1 | 1 GiB | 5 GB, expandable | Production with high availability (Valkey: one primary and two replicas; RabbitMQ: 3 nodes) | E-commerce orders, file generation, integrations |
queue-premium | 2 | 2 GiB | 5 GB, expandable | More volume, with metrics | Many workers and producers, message spikes |
queue-enterprise | 8 | 8 GiB | from 20 GB, expandable | Large message volume | Platforms with intense asynchronous processing |
The vCPU and RAM in the table apply to each instance, and the plan price is also per instance. From queue-basic up, the queue runs 3 instances, so you pay 3 times the plan price, and additional storage is charged on each of the 3 copies. When you create the queue, the Console shows the price per instance and the total in the Cost Summary.
How to choose: from queue-basic up, the queue has high availability, which matters when losing the processing of a message affects the customer. If messages pile up because workers cannot keep up, add workers before considering a larger plan for the queue. On both engines, the queue changes plan among Basic, Premium, and Enterprise; Starter (1 instance) and Free do not change plan, so if the queue needs high availability, start on Basic.
Frequently asked questions
Can I use the same service as a cache and as a queue?
It is not recommended. Cache and Queues have different persistence and eviction settings. Use a Cache for disposable data and a Queue for tasks that must not be lost.
How many free services can I create?
Free Cache and Queues are limited to 1 per profile in each organization; free Queues are Valkey only. See Limits and quotas.
Next steps
RabbitMQ® is a trademark of Broadcom, Inc. in the U.S. and other countries. Zenifra is not affiliated with or endorsed by Broadcom.
Last updated on
Which database plan to choose
What each Zenifra plan for PostgreSQL, MariaDB, Valkey Key-Value and Analytics is for, with examples of uses that make sense on each one.
Quick Start — Node.js
Publish a working Node.js application on Zenifra with Express, GitHub, configured port, and deployment validation.