What is the MOST likely reason for this occurrence?

By:
In: DBS-C01
Tagged:
With: 1 Comment
An online advertising website uses an Amazon DynamoDB table with on-demand capacity mode as its data store.The website also has a DynamoDB Accelerator(DAX) cluster in the same VPC as its web application server.The application needs to perform infrequent writes and many strongly consistent reads from the data store by querying the DAX cluster.During a performance audit, a systems administrator notices that the application can look up items by using the DAX cluster.However, the QueryCacheHits metric for the DAX cluster consistently shows 0 while the QueryCacheMisses metric continuously keeps growing in Amazon CloudWatch.

What is the MOST likely reason for this occurrence?

A VPC endpoint was not added to access DynamoDB.

Strongly consistent reads are always passed through DAX to DynamoDB.

DynamoDB is scaling due to a burst in traffic, resulting in degraded performance.

A VPC endpoint was not added to access CloudWatch.

Explanations:

A VPC endpoint is not required to access DynamoDB. DynamoDB can be accessed without a VPC endpoint if it is configured with an internet gateway or Direct Connect.

Strongly consistent reads bypass the DAX cache and are directly fetched from DynamoDB. This would cause QueryCacheMisses to increase and QueryCacheHits to stay at 0.

DynamoDB scaling due to burst traffic does not explain the continuous increase in QueryCacheMisses. The issue is related to the read consistency, not traffic load.

A VPC endpoint for CloudWatch is unnecessary to access CloudWatch metrics. This option is unrelated to the DAX query cache behavior.

2025-10-02

1 Comment

  1. Terry
    Author

    I map out that the answer is:
    Strongly consistent reads are always passed through DAX to DynamoDB.

Leave a Reply to Terry Cancel reply

Your email address will not be published. Required fields are marked *

19 + 18 =