Hey folks,
I need help with a strange observation...
# supporth
Harish
03/18/2024, 6:50 PMHey folks,
I need help with a strange observation. Here is the context:
I built a Golang application using Gin, deployed it to e2 8core 16gigs Compute instance on GCP.
I built a simple endpoint that makes 2 database query calls and returns the response. Its a Get Todos API endpoint. Im attaching the trace breakdown of the request along with the Postman API call.
TAT on Signoz: 18.29ms
TAT on Postman: 515ms
Compute instance location: us-central1 Iowa
My location: Bangalore, India
I tried making the curl request through gcloud CLI which was also provisioned in Iowa the latency was 400ms, which is still higher.
Anybody has any idea how to reduce the latency here?
Thanks in advance
Harish
03/18/2024, 6:52 PMcurl request breakdown(from my machine)
Copy code
time_connect: 0.210479s
time_appconnect: 0.000000s
time_pretransfer: 0.210531s
time_redirect: 0.000000s
time_starttransfer: 0.423873sHarish
03/18/2024, 6:53 PMcurl request breakdown(from the server itself using localhost)
time_namelookup: 0.000657s
time_connect: 0.000786s
time_appconnect: 0.000000s
time_pretransfer: 0.000822s
time_redirect: 0.000000s
time_starttransfer: 0.005098s
this matches with what OTEL is measuring
a
Ashu
03/19/2024, 6:24 AM
@Vishal Sharma can you check?
v
Vishal Sharma
03/26/2024, 12:04 PMHey @Harish
Are you using nginx or other service inbetween which might be causing high latency? You can instrument nginx as well to check.
(Also I am not sure how we can help you to reduce latency of your application. SigNoz helps you to identify latency of request inside your application which looks it’s doing fine.)
h
Harish
03/26/2024, 12:39 PMHey Vishal, yes I'm using nginx. But for the single request test, I directly hit the application port. I think it's purely network latency.
Yeah I'm aware, the issue is specifically not related to signoz 😅
👍 1
6 Views