https://www.growthbook.io/ logo
w

wonderful-branch-2202

02/28/2023, 5:19 AM
Hey Team, Is there an example usage of FeatureMap available for reference?
f

fresh-football-47124

02/28/2023, 5:20 AM
which language?
its more of a construct than a method
w

wonderful-branch-2202

02/28/2023, 5:31 AM
Got it, was stuck on this part, so graham, I am hosting this locally and have created these features in the growthbook UI, I would have to pass this as feature map inspite of having it configured in the local growthbook container or am I missing some default functionality ?
I am trying to do this in golang
f

fresh-football-47124

02/28/2023, 5:31 AM
there is an API end point to fetch this map
you can go to the SDK section in the feature flags
when you make a new SDK connection you’ll see the full instructions
w

wonderful-branch-2202

02/28/2023, 5:33 AM
Great, thank you so much
Also, Graham a quick question, for A/B testing if we have spilt the traffic, how can we test this component in terms of metrics, say if we toggle the feature among set of users
For A/B testing the button is not toggling what am I doing wrong? Default value is what the experiment is running on
@fresh-football-47124 Can you please help me this? Default value is the only one being picked up. I have configured the key in UI and splitting traffic 50/50. I just want to understand if my understanding is inline with the feature. By splitting this traffic in half we mean that randomness of the feature is 1/2. For some reason I am getting the same feature for the experiment.
f

fresh-football-47124

02/28/2023, 4:50 PM
hi Shruti - can you share the override rule?
like a screen shot?
w

wonderful-branch-2202

02/28/2023, 4:54 PM
This is the feature map that I am using { "enableskipquestion": { "defaultValue": true }, "chatid": { "defaultValue": false } }
f

fresh-football-47124

02/28/2023, 4:56 PM
okay, so you have two experiments there
users will never reach the second experiment, since you are including 100% of traffic in the first experiment
What is the COMPLETED user attribute?
w

wonderful-branch-2202

02/28/2023, 4:57 PM
That is one of the state of chat by the user
I disabled one of the experiment as well but the toggle did not work for the same, but let me cross verify once again
f

fresh-football-47124

02/28/2023, 5:00 PM
check the default value
as well
it shouldn’t matter for an experiment rolled out at 100%
w

wonderful-branch-2202

02/28/2023, 5:01 PM
Sure running the experiment now
Current experiment with feature map { "enableskipquestion": { "defaultValue": true } }
This did not solve the issue, Am I in the correct direction?
f

fresh-football-47124

02/28/2023, 5:07 PM
is this a web test?
w

wonderful-branch-2202

02/28/2023, 5:09 PM
Yes, Web test being targeted by back end
f

fresh-football-47124

02/28/2023, 5:12 PM
So that parameter “COMPLETED” - I have a feeling you don’t mean to use it like that
it sounds like you want to test users who have completed some action
in which case you’ll want to use that as a targeting condition, rather than the SPLIT condition
the SPLIT attribute is used for the assignment
w

wonderful-branch-2202

02/28/2023, 5:16 PM
Ok let me check for this one, thank you so much
Also Graham is Mongo data source not available as of now ?
f

fresh-football-47124

02/28/2023, 5:24 PM
we haven’t had that request currently - you’re storing your event data to Mongo?
w

wonderful-branch-2202

02/28/2023, 5:25 PM
Yes
I will probably store it in different db instance otherwise
f

fresh-football-47124

02/28/2023, 5:31 PM
what event tracker are you using to get your event data to Mongo?
w

wonderful-branch-2202

03/01/2023, 5:12 AM
We are using state machine to track the events
5 Views