Output requests that are over a configurable threshold

[tsenart]

Output requests that are over a configurable threshold

[tsenart]

Hello Sebastian. This is a duplicate to the only current open issue #19. I have to set apart some time to look into it. Thank you for reporting!

[spier]

Thanks for the reply :)

I saw the other issue but I cared less about potential timeouts but rather about slow responses, which is why I filed this as a suggestion. So the two issues might be related but not the same I think.

[tsenart]

Well, a client side timeout is the same as setting a time threshold for a request. Where exactly do you see the difference? If a flag is provided to specify a timeout for each request, and the server takes longer than that, then you will certainly see that on the errors section and on the plot as well.

[spier]

In my suggestion I actually just wanted to visually separate requests that are above and below the threshold. I don't want to force the requests to time out. The advantage is that one would still know how long the requests actually took that were above the threshold.

Use case could be:
Identify all API calls that are above a satisfactory response time. Then debug them one by one.

[tsenart]

You can just look at the plot generated by vegeta report and see which requests are above a certain threshold. For debugging them, you'd have to match up the approximate times of those requests with your server logs.

Assuming the latency is not in the network, but on the server creating the responses, I would recommend to actually just log requests that take longer than the threshold you have in mind. Everything is just clear that way and easier to inspect and correlate.