Support a HEAD request to the API endpoints for latency testing

[bobwith2bees]

Hello,

I am writing status monitoring scripts for the Rarible protocol to alert me when the services have been impacted. I know the team is working on status page efforts - this would likely be useful in that effort as well.

This request is for a simple http HEAD request endpoint that I can call to test TCP connectivity and latency with a tool like httping. It would need to be on every host that provide the APIservices (https://docs.rarible.org/api-reference/)

httping -l -i 10 -GsYK https://www.yahoo.com

[eduardstal]

Hey Bob!
I've added this to our backlog, we should also discuss this feature request in the next developer call.

[bobwith2bees]

There are many places an API call can get delayed. Supporting an empty HEAD request can separate out the following possible causes of latency and errors:

• DNS Lookup
• IP routing
• Load balancer traversal
• Server TCP connection delay (3way handshake etc.)

Then you can test the actual API latency by maintaining the same TCP connection open (created by the HEAD request) then make your API call to a known IP address/endpoint.

If I complain about an API being slow, the first things you would want me to eliminate can be covered by the HEAD request. Otherwise you start the "I don't see the latency on my end" back and fourth.