I have an AWS Lambda, based on typescript, that calls SQL Server (located on an AWS EC2 instance), using node-mssql library for this purpose.

I would like to trace the request and find the bottleneck. In my code I can cover the calling method and calculate the time.

So, the request can be split into the following parts:

1. init calling on Lambda's side
2. the part from lambda to EC2 (network delay)
3. the "incoming" part inside EC2 refers to the time when the virtual server calls SQL Server internally
4. T-SQL query execution time
5. the "outcoming" part inside EC2
6. the part from EC2 to lambda (network delay)
7. return the result to the calling code

I'm confident about step #4; I profiled the SQL query time, and it's quite small.

So, the problem lies in another step or steps.

I'm exploring X-ray and have configured it for calling S3 as an example, but there seems to be no way to profile the node-mssql library (perhaps another library could work?). Therefore, I'm unsure how to implement this in practice.

My idea is that we can at least monitor the network parts (#2 and #6) because I don't see any barriers to observing network requests (but I did not find how to do it using X-ray). Additionally, since EC2 is part of the AWS family, it probably could provide us with some helpful insights or tools.

Any ideas and suggestions? :)

Thanks