Checkin takes > 1 min with ~2.5 million records

Environment:
AWS Aurora
Engine: Aurora PostgreSQL 9.6.3
DB instance class: db.r4.large
vCPU: 2 vCPU
Memory: 15.25 GiB
EBS Optimized: 400 Mbps
Network Performance: 10Gbps

AWS RDS
Engine: RDS PostgreSQL 9.6.6
DB instance class: db.r4.large
vCPU: 2 vCPU
Memory: 15.25 GiB
EBS Optimized: 400 Mbps
Network Performance: 10Gbps

Steps to reproduce:
Inventory App:
Copy barcode whose status is "available" from HoldingsRecord after selecting any Title in Inventory App

Checkout:
Click Patron look-up and select Administrator or any User available, then in right pane enter valid barcode to check out for example "5705848762"(copied from above) and click Enter button

Checkin:
Use the same barcode from Checkout for example "5705848762" to check in. Checkin takes > 1.5 minutes. Please see attached snapshot in the attachment section.

It queries circulation/loans endpoint to get userid and
then it queries the users using that userid to verify if the item is loaned then it queries holdings-storage/holdings endpoint to check in.

Please see attached log retrieved from Chrome developer tool.

A possible solution to fix this issue:
Why not make a direct call using holdings record Id if it is available? - this should take around 0.5 seconds
http://tmp-lb-pub-185231094.us-east-1.elb.amazonaws.com.:9130/holdings-storage/holdings/4c415313-a5b8-49b5-accd-c6bcf156fff0

vs

Passing id as a query - takes around 1.5 minutes
http://tmp-lb-pub-185231094.us-east-1.elb.amazonaws.com.:9130/holdings-storage/holdings?query=%28id%3D%224c415313-a5b8-49b5-accd-c6bcf156fff0%22%29