Often when asked about scaling cache in a microservice architecture, novice developers start talking about distributed cache which can be an overkill in most cases. So the following are the steps to be followed according to me.
- Do not have any cache
- Have embedded cache (i.e. the application caches the database information itself)
- Have a separate single node cache
- Have a multi-node distributed cache that uses consistent hashing to choose the cache node
- Have an autoscale cloud cache like Amazon ElasticCache