concern over a question – AWS RDS slow response time != replica lag

[samu-2]

Dear TutorialsDojo,

I have a concern about a potential error in one of the questions of the “AWS Certified Database Specialty Practice Exams 2020” course.

The question looks like this, where I have highlighted the problematic part:

A company has an enterprise application hosted in an Amazon ECS cluster and uses Amazon RDS for MySQL database with a read replica. The application uses the read replica to generate business reports and avoid performance issues on the primary DB instance. However, the users noticed that the queries to the read replica have a slow response time.

Which of the following is the MOST likely root cause of this issue?

– The replica lag metric has reached a value of 0.

– The change data capture (CDC) option is enabled on the primary DB instance for continuous data replication.

– There are long-running queries on the primary DB instance. (this one is the correct answer according to the course)

– The replica DB instance class is higher than the primary DB instance.

The problem with this question is that in reality, long running queries, in my opinion, should not cause slow response time on the read replica. They will cause replication lag but not slow response time.

Therefore I think the question and the answer are incorrect and misleading.

Do you agree?

• This discussion was modified 3 years, 5 months ago by  samu-2. Reason: clarified which one is the correct answer according to the course

[Joboy-Pineda-TD]

Hello samu-2,

Thank you for the post. I agree with your point that the slow response time issue does not directly relate to the replication lag. Long-running queries in the primary instance could potentially impact the replication lag between the primary instance and the read replica but do not necessarily slow down the performance of the read replica directly.

What we intend to highlight in this scenario is the relationship between the primary RDS instance and its replica. To prepare for the exam, it would help the candidate to understand what happens during the replication process and what you should take note of whenever you observe that the data takes a while to synchronize.

Happy to help if you have any other concerns or clarifications. Good luck with passing your AWS Database Specialty Exam.

Joseph Pineda

[samu-2]

Hi Joseph,

thank you for the reply!

In that case, I would suggest altering the wording of the question a little, to reflect the replica lag, as opposed to slow response time. When we say “slow response time”, I think most people think of latency and not staleness of data. So, just for complete correctness, I would suggest anything expression from the following:

[…] However, the users noticed that the data lagging behind / stale / not up to date.

Have a great weekend and again thank you for your reply.
Best regards,

Samuel