What should be done to address this issue and improve performance?
Scale the cluster by adding additional nodes
Scale the cluster by adding read replicas
Scale the cluster by increasing CPU capacity
Scale the web layer by adding additional EC2 instances
Explanations:
High eviction count suggests that the cache is unable to hold all the data, likely because of insufficient memory. Adding additional nodes increases the memory capacity, which can reduce evictions and improve performance.
Read replicas are useful for offloading read queries in databases, but Memcached is a caching layer, not a database, and does not benefit from read replicas. This option does not address eviction issues.
Scaling the cluster by increasing CPU capacity does not address the issue of memory capacity, which is the root cause of high eviction count. CPU capacity may not be the limiting factor here.
Adding more EC2 instances may help distribute load across servers, but it does not address the root cause of high eviction count in Memcached, which is related to insufficient memory in the cache.