Research Elasticache parameter settings and service offerings, and work with cloud.gov to make adjustments #1480
Description
We've recently learned that we can make some adjustments (and/or come up with a whole new cluster configuration) to the AWS ElastiCache service within cloud.gov, and there are several things we ought to factor in to make some adjustments.
The most important change we ought to make and request is adjusting the maxmemory-policy to make sure old keys get cleared out when space starts to become an issue. A different cluster mode might also be necessary.
Outside of that, there are several things we need to research:
- Memory management: https://docs.aws.amazon.com/AmazonElastiCache/latest/dg/redis-memory-management.html
- ElastiCache parameter groups: https://docs.aws.amazon.com/AmazonElastiCache/latest/dg/ParameterGroups.Engine.html#ParameterGroups.Redis.4-0-10
- Configuration options (in case there's something else that might suit us better): https://aws.amazon.com/elasticache/pricing/
- Redis key eviction: https://redis.io/docs/latest/develop/reference/eviction/
- ElastiCache cluster modes: https://docs.aws.amazon.com/whitepapers/latest/database-caching-strategies-using-redis/redis-cluster-modes-enabled-and-disabled.html
- More info on cluster mode management: https://medium.com/@anil.goyal0057/understanding-elasticache-for-redis-cluster-configurations-and-scaling-options-46c94ade9f8e
This is the information cloud.gov had sent us in response to using up all of the memory in our Redis instance previously. Let's go through this information and determine what our next steps ought to be.