Data Caching
There are two main types of caching used within Node.js applications. These include:
- in-process
- cluster wide
Recommended packages#
| Caching type | Recommended client |
|---|---|
| in process | lru-cache |
| cluster wide | ioredis |
In process#
Recommendation is to use lru-cache which is broadly used within the
the teams organizations and in the JavaScript ecosystem with > 28 million weekly
downloads.
Cluster wide#
For cluster wide caching with Redis, the recommendation is to use ioredis.
There are 2 good options with no clear technical winner for redis, but the teams
organizations have experience in production with ioredis so
that is why it is recommended over redis. Depending on the server you are
using there may be features which are only available in a server specific client and
if you need those features then using the specific client makes sense insteda of ioredis.
Guidance#
- Consider your call behaviour to ensure you need to cache a specific piece of data at all
- For cluster wide caching, include a version prefix as part of keys from the start, otherwise you will encouter problems when you try to update the caching schema