I am looking for guidance in terms of what the best approach is for implementing unit tests with the latest SDKs. In general it seems there needs to be support for the following. • a flag or something to enable a local mode during initialization (where it doesn't go out over the network) • a way to explicitly override a specific feature flag with a unit test (used during the setting up initial state portion of the test) • a way to clear all feature flag overrides (used in a hook before each test case to make sure we don't have state bleeding across tests) I am particularly interested in the Javascript, Ruby, and Flutter SDKs. Reading through the Javascript docs, https://docs.growthbook.io/lib/js, I don't see anything that would address the above needed functionality. How are people writing test cases for the on & off cases when a feature toggle is in play. Please provide any information or guidance. Reading through the Ruby docs, https://docs.growthbook.io/lib/ruby#dev-and-qa-helpers, it seems to provide the

gb.forced_features =

assignment operation which seems like it would facilitate two of the 3 needed features listed above. But it doesn't indicate anything about having it not reach out to the network with a local mode setting. Again any guidance around how people are testing the on & off use cases when a feature flag is involved would be helpful. Reading through the Flutter docs, https://docs.growthbook.io/lib/flutter and https://github.com/growthbook/growthbook-flutter, it isn't clear how we should be writing our unit tests when a feature flag is in play either. Again please provide any guidance.

Hi! Is it possible to use variables (attributes) inside JSON type flag? Currently, we have own feature flags implementation, and it may be helpful to use i.e. $USER_ID variable to customize some links inside app.

Hi, any idea on the steps to implement feature flags on next js 14 especially wrapping the

<GrowthBookProvider></GrowthBookProvider>

tag on the app

Hi folks, I have one feature flag using the Experiment at Override Rule. My Experiment is split into 2 kinds: one is control, and one is treatment with the setting of 50-50 (I am using a list that has 5 IDs). But when I check in the list of IDs I see only one ID is in the treatment group. All of the 4 remaining IDs are in group control? How can I verify this issue?

Hi! Just wondering if there are any other BitBucket Pipelines users out there who've looked into setting up the GrowthBook's Code References CLI tool? Seems like something that a pipe could be set up for with (relative) ease and wanted to make sure there wasn't one floating around before I sank time into it...

Hey team I’m not sure where to put this question. I’m adding new attribute

country-code

as per document https://docs.growthbook.io/features/targeting#country-code-targeting But I’m not able to get

Change the String Format to ISO Country Code (2 digit)

option. Is this included only in paid version of GrowthBook?

Hi Growthbook Team, We have created a feature key ("dm_feature_key") of the feature type JSON. We are accessing it using the code below:

function get_feature_values(){
    const value = window._growthbook?.getFeatureValue("dm_feature_key", {} );
    console.log(value)
    return value;
}

We have one control and one variation with a traffic split of 5% and 95%, respectively. Often we receive the JSON, but most of the time we get an empty dictionary instead of the JSON defined in variation 1.

Hi! My setup for feature toggles requires that I have quite a lot of environments. This is a POC that I've setup and it's great but I'm concerned that the environment list is going to grow and it's going to become unmanageable. Has anyone else had to do anything similar? Are there any plans to allow for creating different views for environments? I'd love to be able to filter environments just like I can filter/search features and tags.

Hey, all. Has anyone gotten this feature working? Each time I've enabled it, I've had my JS SDK up-to-date, but it doesn't work. The first time I tried, I had a constant stream of errors show up and no feature flags worked. I waited a month and now when I just tried, there are no errors, but the flags that should turn on for the group I'm in don't turn on. Once I turn the feature off and refresh my site, the features toggle as expected. Also, the payload from GB looks good with the groups broken out and given unique IDs.

Hi GrowthBook Team 👋 I would like to use the Date format for a string attribute and I'm wondering which date format is expected. Can you please clarify? ☺️

Hi GrowthBook Team, I am using Android Kotlin SDK. However, it is not returning any features. How can I confirm wherther SDK is connected properly or not?

Hello, I can't find any information. I'm using nextjs 15 and app routing. Should I create 2 growthbook instances for client and server?

Hi, I just set up GrowthBook with BigQuery successfully and tested a featureflag. A 50/50 boolean test and I verified two users got different values. But where can i actually see the data on how the featureflags and experiments are going? Like amount of users in that's seen/used the flag. Should I also be able to see the data from these in BigQuery? There are no rows in the data-set that I connected too.

Anyone who could help me out? Pease

Hi. I want to use the same feature in 2 different experiments in parallel. But all time i receive the value from "the first" experiment. Is any option, to receive feature value specific to experiment_id or key, its so generic for me to have getFeatureValue("MainBtnColor") forExperimentKey("experiment_2")

Hello! I am just new to growthbook and trying something new in the feature-flags however, i feel struggled when using saved group, as the default 'id' is in string data type, i can use 'id' as attribute key for the saved group when i create a new id with integer type, it seems not work may I ask if this is a normal phenomenon?

Hello everyone! I'm hoping to get some help on a deployment issue I'm facing while setting up a self-hosted GrowthBook instance on AWS. High-Level Problem: The application deploys and the UI loads, but immediately shows an error, because the initial

GET /api/init

call consistently fails with a

401 Unauthorized

error. The crucial finding is from the request headers for `/api/init`: there is no

Authorization: Bearer ...

header present, but there is a

cookie

header containing a

gb_session_id

. My Environment: • Platform: AWS ECS (on Fargate). • Database: MongoDB running as a sidecar container in the same ECS task. • JWT_KEY: Defined a long random string • Networking: An Application Load Balancer (ALB) with HTTPS listeners. • Image:

growthbook/growthbook:latest

When I clone the Growthbook feature flags example to Vercel and then set my

GROWTHBOOK_CLIENT_KEY

and

FLAGS_SECRET

then deploy I am seeing this error:

Hello everyone, I need some help on this - I'm facing an issue where when I click on any of the features I created, I'm getting an error. Kindly help me on this

Hi dev team. Not a bug or question, just a casual suggestion to have something e.g. a confirmation popup before really toggling the flag, to prevent doing that by mistake/misclick. This happened to me yesterday, on production (fortunately no major impact). People asked me why I turned it ON but I had no idea, I just visited the page and maybe somehow clicked the toggle 😢. FYI, our company's version is Build: 3.5.0+4470a21 (2025-03-04) Thanks for your attention.

Will these two rollout rules in the same feature flag target the same 15% of users or different 15%s?