Monitoring ElastiCache for Performance and Availability

Question 1

ElastiCache supports two engines

Answer 1

Memcached and Redis

Question 2

Monitoring Metrics

Answer 2

• -CPU Utiliazation
• -Evictions
• -CurrConnections
• -Swap Usage(Memcached)

Question 3

CPU Utilization

Answer 3

CPU host-level metrics

Question 4

CPU Utilization – Memcached

Answer 4

• • multi-threaded]
• • can handle loads of up to 90%
• • above 90% becomes a problem
• • Solution: increase the size of the node or scale out by adding more nodes;

Question 5

CPU Utilization – Redis

Answer 5

– calculate the threshold:(90/# of CPU cores)
–Solution: for read-heavy workloads, increase the number of read replicas
for write-heavy workloads, use a larger cache instance.

Question 6

Evictions

Answer 6

happen when a new item is added but there is no more memory space. an older item must be deleted to make space.

Question 7

Evictions – cont.

Answer 7

• • can be a caching technique used to make sure you do not run out of memory
• • if items get evicted too frequently, it defeats the purpose and will decrease performance
• • CloudWatch alarms can notify you of a certain threshold.
• • Memcached solution: increase instance size or add nodes to your cluster
• • Redis solution: increase the node size.

Question 8

Current connections

Answer 8

an increase in currconnections could indicate a larger problem with your application

• • the application may not be releasing connections
• • Choose a threshold based off of your application requirements.

Question 9

Swap Usage(Memcached)

Answer 9

caused because the memory allocated for connection information and other overhead items gets mazed out
– should stay at 0, and not exceed 50MB.
– affects performance and should be avoided.
Solution: increase node size
– increase our ConnectionOverhead parameter value(this will decrease memory available for caching data).