How can I show the latencies of the internal computations of a service in locust?

Is there a way to show the latency statistics (and hopefully graphs) of the internal computations of a service in the locust web interface during the test?

I have a service which internally realizes several computations. I would need to perform a load test and benchmark the times of each of these internal computations. Something like:

/compute:

• Computation A -> Xms average, Yms median, Zms max etc.
• Computation B -> Xms average, Yms median, Zms max etc.
• Computation C -> Xms average, Yms median, Zms max etc.

However in locust I can only see the overall time statistics of the endpoint (/compute in this case).

I am right now returning the latencies of each of the computations in the response. I have checked the docs but I have not found a way to do show the statistics of those numbers during the test in the locust web interface.

The only workaround I have found is saving the responses into a file and computing the statistics separately.

Is there any way to do it? Or any other better solution?

Thank you so much in advance

Yes, there is a way to do it. What you're looking for is Extending the Web UI.

> As an alternative to adding simple web routes, you can use Flask
> Blueprints and templates to not only add routes but also extend the
> web UI to allow you to show custom data along side the built-in Locust
> stats. This is more advanced as it involves also writing and including
> HTML and Javascript files to be served by routes but can greatly
> enhance the utility and customizability of the web UI.
>
> A working example of extending the web UI, complete with HTML and
> Javascript example files, can be found in the examples directory of
> the Locust source code.

Specifically, this is the directory with the relevant example of extending the web UI. It's a bit more involved to extend the web UI than it is to just add your custom route, but if you're familiar with the technologies it shouldn't be too bad. You can add your own tab to the UI and show whatever you want on it. You can even create a table and a chart in the same style as the main Locust stats tables and chart if desired.

A quick workaround might be firing your own request event for each computation so they show as different endpoints in de web UI.

  request_meta = {
        "request_type": "my-custom-type",
        "name": "my-custom-name",
        "response_time": <calculated-time>,
        "response_length": 0,
        "exception": None,
        "context": None,
        "response": None,
    }  
  env.events.request.fire(**request_meta)

Writing a wrapper might result in a cleaner code too.