0

My initial understanding of my read-replica is that it would be less busy than the primary. After all, the primary is processing all the reads, inserts, updates and deletes which are supporting my transaction processing. All the read-replica would be doing is serving the interactive demand for reports.

But after thinking about this, and working with the read-replica, I think I'm completely wrong about this. The read-replica is, I suspect, supporting a greater load than the primary because the read-replica needs to support all of the aforementioned reads, inserts, updates and deletes PLUS the interactive demand for reports.

Am I right about this? Does synchronization utilize the vCores allocated to the replica? Or is the synchronization between the primary and the replica performed at a lower level (filesystem or disk IO), which would not be utilizing the vCores allocated to the replica?

Thanks for your advice!

1 Answer 1

0

Actually, it appears that my initial understanding was correct. Here are the side by side (top / bottom) comparisons of the activity profile for the primary and replica databases. Clearly, the replica is utilizing much less CPU resources than the primary.

enter image description here

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.