Which of the following actions would allow the database to handle the month-end load with the LEAST impact on performance?
Pre-warming Elastic Load Balancers, using a bigger instance type, changing all Amazon EBS volumes to GP2 volumes.
Performing a one-time migration of the database cluster to Amazon RDS, and creating several additional read replicas to handle the load during end of month.
Using Amazon CloudWatch with AWS Lambda to change the type, size, or IOPS of Amazon EBS volumes in the cluster based on a specific CloudWatch metric.
Replacing all existing Amazon EBS volumes with new PIOPS volumes that have the maximum available storage size and I/O per second by taking snapshots before the end of the month and reverting back afterwards.
Explanations:
Pre-warming Elastic Load Balancers and changing instance types would not address the database’s heavy I/O bottlenecks directly. GP2 volumes may not provide sufficient performance for heavy workloads, especially during month-end.
Migrating to Amazon RDS and adding read replicas allows for better scalability and load distribution, which directly addresses performance issues during peak traffic periods. This solution minimizes the impact on the existing setup while enhancing performance.
While dynamically changing EBS volume types or IOPS can help, it may not be fast enough to handle sudden spikes in demand during month-end reporting, leading to potential delays and performance issues.
Replacing EBS volumes with PIOPS volumes may improve performance, but taking snapshots and reverting back could cause downtime or data inconsistency issues, negatively impacting performance during critical periods.
If I’m correct, the answer is:
Performing a one-time migration of the database cluster to Amazon RDS, and creating several additional read replicas to handle the load during end of month.
In my experience, the answer is:
Performing a one-time migration of the database cluster to Amazon RDS, and creating several additional read replicas to handle the load during end of month.