Locust does not count RPS if all requests fails

[frenetic]

Description of issue / feature request

Im testing an endpoint, which is expected to returns HTTPS 403 (99.9% of the scenarios I have in my file).
Im doing this to test how this endpoint deals with its DB, Redis and 3rd party auths.

Expected behavior

I was expecting to see the correct number of requests per second

Actual behavior

Locust says it is doing 0 requests per second.

Actually, if we see through the next chart - even though Locust is displaying the correct AVG response time, it only display the RPS for successful requests.

Environment settings (for bug reports)

• OS: Ubuntu 16.04 x64
• Python version: Python 2.7.12
• Locust version: 0.8.1

Steps to reproduce (for bug reports)

I cant share my locust.py, since it holds sensitive data.
However, just write a test for an endpoint that always returns 403 - no matter what.

[myzhan]

It's a feature, you can call events.request_success.fire() directly to work around.

See https://docs.locust.io/en/latest/testing-other-systems.html

pseudo code

response = requests.get(endpoint)
if not response.ok:
    events.request_success.fire(request_type="http", name="fail", response_time=100, response_length=10)
else:
    events.request_success.fire(request_type="http", name="success", response_time=100, response_length=10)

[frenetic]

I think I might be stupid.
Thank you for your help.

[heyman]

@frenetic No worries :). We've chosen to only include requests that's considered "successes" in the RPS, but I think one could definitely make a case for including all requests that resulted in a response from the web server.