Exam AWS Certified SysOps Administrator - Associate topic 1 question 189 discussion

I'll give a tip : read latency : read replica write latency : multi-master

The key word is "dowtime", so I vote for A. B is out because this changes needs a dowtime for business

Key word is increase in select latency is causing the high memory demand. Select is used for query purpose only so a extra read replica will solve this

'Select' latency - read-only workload causing high memory consumption and latency. Therefore Read replica - A

Keyword is select latency, which means read operations. So adding Read Replica will reduce Read Latency and also RAM consumption, load on the Primary.

ans. A

A. Add an Aurora Replica to the DB cluster. To improve the performance of the Amazon Aurora DB cluster with minimal downtime and no data loss, adding an Aurora Replica is a suitable option. Here's why: Adding an Aurora Replica (Read Replica): This helps offload read traffic from the primary (writer) node to the replica (reader) node. It allows the read workload to be distributed and can improve overall performance. Minimal Downtime: Adding a read replica is a non-disruptive operation and can be performed with minimal downtime. It involves creating a read replica from the existing primary instance. No Data Loss: Creating a read replica does not result in data loss. The replica is initially synchronized with the primary instance and continues to replicate changes as they occur. -- Options B, C, and D involve more significant changes or potential downtime:

Keyword is downtime so it is A

C "If, after investigating your workload, you find that you need more memory, consider scaling up the DB instance class to a class with more RAM." https://docs.aws.amazon.com/AmazonRDS/latest/AuroraUserGuide/Aurora.BestPractices.html

"The change must minimize downtime" <- its all about this quote I agree that C would solve the issue in a better way, but migrating DBs is a operational nightmare.

During sudden traffic surges, the high RAM consumption and increased select latency indicate that the current memory-optimized instance types might not be sufficient to handle the increased load. To improve the performance of the DB cluster, the best approach is to scale up the instance types to larger memory-optimized instances. Option A (Add an Aurora Replica) would help with read scalability and high availability but may not directly address the high RAM consumption and select latency issues. Option B (Modify the DB cluster to convert it into a multi-master DB cluster) doesn't directly address the high RAM consumption and select latency issues either. It would be more suitable for scenarios where you need to improve write scalability. Option D (Increase the disk storage capacity) would only address storage-related issues and would not directly improve the performance of the DB cluster in terms of RAM consumption and select latency.

Not enough info to decide.

Add an Aurora Replica to the DB cluster. For select latency

To improve the performance of the Aurora DB cluster, while minimizing downtime and without losing data, option A - adding an Aurora Replica to the DB cluster, is the best choice. Adding an Aurora Replica would create a new reader node in the cluster, which can help distribute the load during sudden traffic surges, and reduce the select latency. This can also help reduce RAM consumption on the writer node. Additionally, adding a replica can be done with minimal downtime, as it does not require any changes to the application, and data will be automatically replicated from the writer node to the new replica.

Option C, taking a snapshot of the DB cluster and creating a new DB cluster with larger memory-optimized instances, will allow the SysOps administrator to increase the memory capacity of the instances without losing any data. This option also minimizes downtime as the new DB cluster can be created from the snapshot while the original DB cluster continues to serve traffic. Once the new DB cluster is created, the administrator can update the web application to use the new DB cluster and then decommission the original DB cluster.

I say C. The high RAM consumption and increased select latency suggest that the memory-optimized instances in the Aurora DB cluster are struggling to handle the sudden traffic surges. To improve the performance of the DB cluster, the SysOps administrator should increase the memory capacity of the instances. Option A, adding an Aurora Replica, will not increase the memory capacity of the instances, so it is not a suitable option in this scenario. Option B, modifying the DB cluster to convert it into a multi-master DB cluster, may improve write performance, but it is unlikely to improve read performance or memory capacity. Additionally, this option may result in downtime and require more significant configuration changes than other options. Option D, increasing the disk storage capacity of the DB cluster, will not address the high RAM consumption and increased select latency, so it is not a suitable option in this scenario.

Memory Issues and select latency are more related with read process