Hey all, two questions about API/webhooks: 1. Chan...
# announcements
w
Hey all, two questions about API/webhooks: 1. Changing experiments Live and doing other changes to the experiments (in Analysis --> Experiments) doesn't trigger a webhook? Is there a way to make it trigger a webhook? 2. We're using the https://cdn.growthbook.io/config/ API to pull some experiment conf data but I can't find any documentation for that endpoint anymore. Will it still be supported going forward and what's the reason for removing it from the docs?
f
Hi. We changed the experiments section to just be about analyzing results. The actual implementation of experiments is done through feature flags now. The config API endpoint was replaced by the features endpoint.
The old config endpoint still works, but will not be supported going forward.
w
hmm okay, so what's the purpose of the various configurations under experiments now then?
f
The config settings that remain are used for the analysis. We still have some work to do on the UI to make the separation clearer.
w
okay, thank you for the information.
I must say, when using the tool for A/B testing only, the UI is quite counterintuitive and there's too much unnecessary repeating tasks. Some examples: • Setting experiment status in two different places while only one of them has the actual effect • Setting experiment name in two different places • Choosing targeting, tags etc in two different places In those areas, the old setup worked much better but I hope in time the new one will be streamlined as well.
Oh, and is there a way to apply targeting rules similar to "Targeted" feature to an "A/B Experiment" to bucket visitors into the experiment (not a specific variant)? Thanks.
One more. Is it possible to set the "namespace" somewhere from the UI?
f
Yes, we do plan to make the process more streamlined. To answer your questions: 1. For experiment rules in a feature, you can add "Targeting Conditions" that control who is included in the experiment 2. Once you setup namespaces under settings, you can add them to experiment rules (there's an "Enable Namespace" toggle at the bottom when editing the rule)
w
Thanks, this is. very helpful!
Another quick one: can I pass multiple rules into
growthbook.run()
? i.e. sth like this
Copy code
growthbook.run({
    "defaultValue": 0,
    "rules": [
        {
            "variations": [0, 1, 2],
            "weights": [0.334, 0.333, 0.333],
            "key": "rd005",
            "hashAttribute": "id"
        },
        { "condition": { "id": "123" }, "force": 1 }
    ]
});
f
No. The
run
method just takes a single experiment definition. https://docs.growthbook.io/lib/js#inline-experiments
That's an interesting use case that we haven't considered though. Basically evaluating an inline feature instead of an experiment.