Which combination of steps should the company take to meet these requirements?
(Choose two.)
Deploy an ElastiCache for Redis cluster with read replicas and Multi-AZ enabled.
Deploy an ElastiCache for Memcached global datastore.
Deploy a single-node ElastiCache for Redis cluster with automatic backups enabled. In the event of a failure, create a new cluster and restore data from the most recent backup.
Use the default maintenance window to apply any required system changes and mandatory updates as soon as they are available.
Choose a preferred maintenance window at the time of lowest usage to apply any required changes and mandatory updates.
Explanations:
Deploying an ElastiCache for Redis cluster with read replicas and Multi-AZ enabled ensures high availability and fault tolerance. Multi-AZ provides replication across different availability zones, while read replicas can help balance the load and improve read performance, contributing to minimal downtime during maintenance activities.
ElastiCache for Memcached is not well-suited for handling highly available, fault-tolerant architectures that require automatic failover, as it lacks native support for Multi-AZ and read replicas. Memcached is better suited for use cases with simpler data needs and less focus on availability and durability.
A single-node ElastiCache for Redis cluster with automatic backups does not provide high availability or fault tolerance. In case of a failure, a single-node cluster cannot automatically failover, and restoring data from backups may result in extended downtime and inconsistency in the leaderboard data.
Using the default maintenance window for applying system updates can lead to unplanned downtime during peak usage hours. This does not meet the requirement of minimal impact on the availability of the gaming platform.
Choosing a preferred maintenance window during the lowest usage hours ensures that system updates and changes can be applied with minimal impact on player activity, addressing the requirement for minimal disruption during maintenance.
From my point of view, the answer is:
Deploy an ElastiCache for Redis cluster with read replicas and Multi-AZ enabled.