Q: 10
A company hosts a production MySQL database on an Amazon Aurora single-node DB cluster. The database is queried heavily for reporting purposes. The DB cluster is experiencing periods of performance degradation because of high CPU utilization and maximum connections errors. A CloudOps engineer needs to improve the stability of the database. Which solution will meet these requirements?
Options
Discussion
Option A is the best fit here. Adding Aurora Replicas directly addresses connection and CPU issues since reads are offloaded without extra app logic. D sounds good but assumes all reporting can be effectively cached, which is a trap in production. Disagree?
Not sure A's the only way here, D seems like a solid pick if caching reports is ok. The trap is that ElastiCache offloads reads too. Thoughts?
I don't think it's A. D seems like a solid answer since ElastiCache would offload a lot of the reporting queries and help with CPU spikes. If caching logic is easy to add, pretty sure D could work here. Disagree?
C or D, but I'm actually sticking with A. Only Aurora Replicas (A) let you natively scale read workload off the main DB, plus you can auto-scale based on CPU which matches the issue. Not 100% sure if ElastiCache would fit since you'd need to rebuild query logic.
I don’t think B solves the problem, so D. ElastiCache could offload a lot of reporting queries if reporting data doesn't change often, and fallback to Aurora when it's not cached. Pretty sure that's valid for scaling read-heavy loads, but maybe I'm missing something.
Maybe B. Had something like this in a mock, pretty sure splitting clusters between AZs was suggested for heavy read workloads.
Scaling with Aurora Replicas is made for this, so A. Directly adds read capacity and prevents hitting CPU/connection limits. D's fine for caching but you'd risk stale data, which isn't mentioned as ok here. Pretty sure about A, but challenge if you disagree.
A is wrong, A. Official AWS study guide and practice exams recommend scaling reads with Aurora replicas for this scenario.
Probably A. D looks tempting because of caching but I think that's a trap unless they say reporting can tolerate stale data. Aurora replicas scale reads without adding consistency problems, so A fits best here.
A seems right to me for scaling reads directly in Aurora. Replicas can take that reporting load, matches what AWS recommends. Not 100 percent but leaning this way unless there's a catch with the reporting requirements. Agree?
Be respectful. No spam.