Managed Redis
Redis as a cache, and Redis as a queue.
Version 7, provisioned with credentials and TLS, sitting on the private network beside the application that needs it.
Redis usually arrives for one of two jobs and ends up doing both: holding things that are expensive to compute, and holding things that are waiting to be processed. The two have very different consequences when the instance restarts, and it is worth knowing which one you are relying on.
A cache can lose everything and the application recovers. A queue that loses everything has lost work. Dockup provisions Redis 7 with daily encrypted backups, so the second case is survivable.
- Version
- Redis 7
- Connection
- Resolved by name over TLS inside your workspace
- Public access
- None by default
- Typical use
- Session storage, response caching, rate limiting, background job queues
- Backups
- Daily, encrypted, seven retained
- Cost
- Metered CPU and memory — a cache is usually the cheapest thing in the workspace
Cache and queue on the same instance
It is common, and usually fine, to run both on one Redis. It stops being fine when the cache grows enough to trigger eviction, because the eviction policy does not know which keys are jobs.
If both jobs matter, separate them. Two Redis instances cost very little at this size, and the failure mode you avoid is the one where a traffic spike silently deletes queued work.
Other guides in this section
- Managed PostgreSQLManaged PostgreSQL 17, 16 or 15 with credentials injected, TLS by default and no public hostname. Daily encrypted backups.
- Managed MySQLManaged MySQL 8.4 or 8.0 with credentials injected, TLS by default and no public hostname. Daily encrypted backups.
- Managed MongoDBManaged MongoDB 8 or 7 with credentials injected, TLS by default and no public hostname — the default that prevents exposure.
Read next
- Redis as a cache and as a queueRedis cache and queue patterns on Dockup: provision managed Redis, connect privately, define failure behavior, avoid data loss assumptions, and monitor usage.
- Scaling a database that is already liveFrom vertical scaling to sharding: learn the strategies that will keep your database performant as you grow.
Questions this page gets asked
Can I use Redis as a job queue?
Yes, and most applications do. If the same instance is also a cache, watch the eviction policy — it does not distinguish cached values from queued jobs, so a memory spike can delete work.
Is the data persisted?
Yes, with daily encrypted backups retained for seven days. A cache does not need that; a queue very much does.
What does Redis cost?
Metered CPU and memory at the standard rates. A small cache is typically a couple of dollars a month, and usually the cheapest thing in the workspace.
Start on the free plan.
Ten dollars of infrastructure credit is already there, and no card is required.