Hi all, I’m new here. Do you guys have an architectue diagram? I’m particularly interested in how feature flag works in details. And how many environments do you support

Encountering some bizarre behavior with

@growthbook/growthbook-react

. We’re using

nx

as our mono-repo tool, which could likely be a factor. We have a specific

feature-flags

package that houses our

FeatureFlagContextProvider

component which renders the

GrowthbookProvider

. This gets imported and rendered at the root of our

web-app

(separate “package”). If I write some code to

useFeature

in either of these packages we have no problems. We then have a separate

queries

package that houses all of our

react-query

queries for loading application data. I was attempting to use a feature in one of the queries, however, in debugging it was clear that this line in

useFeature

const { growthbook } = useContext(GrowthbookContext)

was the culprit in that

growthbook

was

undefined

. It’s interesting because if I call the function that wraps this logic side by side with

useFeature

in

web-app

, the

useFeature

in

web-app

correctly gets the feature value whereas the one in

queries

is unable to because

growthbook

is

undefined

. Anyone encounter a similar issue ever?

Hi, everyone, this is Abhushan from Alippo, where I work as a Frontend Developer (working with Next.js) and have been using Growthbook for the last 2-3 weeks. I needed some help/ideas regarding a scenario that we are planning to test for a new feature for our audience using either a simple A/B experiment or a gradual percentage rollout, where I need to conditionally hide or show the feature, Let's say I need to A/B test section C (hide/show) on my page which has the following structure: Section A Section B Section C ----> needs to show/hide this section for certain users Section D Section E I had two approaches to solve this but was wondering if I am heading in the right direction here: 1. Since the section sits in between, conditionally rendering this on the client would be a bad idea as it will cause a significant Cumulative Layout Shift (CLS), and the only way to avoid it would be to restructure the page such that I move the section C to the bottom of the page : Section A Section B Section D Section E Section C -----> conditionally rendering this will cause zero CLS? 2. Find out whether or not to show the section on the server-side (Next.js Data Fetching hook getServerSideProps) and then pass the boolean flag as props and conditionally render Section C (since the page will be server side rendered), which should again not cause any CLS....? Also in general wanted to ask how you plan A/B on client-side (maybe based on a certain user attribute that is only available on the client ) without causing any CLS on the page.( I usually play with opacity 0 and 1 if has a small layout, hiding or showing banners on top of page ) which does not affect CLS, but this does not seem scalable for bigger section since it will leds to empty spaces on the page. Thanks! Loking forward to your suggestions.

Hello everyone. Im trialing out growthbook for my current company and I'm having to hack around the fact that (please correct me if I'm wrong) a feature flag's

default value

is shared between environments, instead of being distinct to each. Say I have feature

X

working in

test

under a boolean

on/off

flag, so I just want to enable it in

test

and have the default value as

on

. For

prod

I want to

X

to be

off

for everyone except for some internal users, but if I enable it in

prod

and set the default value as

off

it then affects the

test

env. My current hack is to enable the feature in all envs with default value as

on

, then in

prod

adding an override clause that turns on the feature if the userId matches some condition. If there are distinct default values for each env then they are isolated and you won't risk accidentally breaking one when configuring the other. Plus I can see that

override rules

are already env-scoped. So I was just wondering if there are specific reasoning behind making

default value

a global config. I'm very new to feature toggling in general any suggestion/criticism is more than welcome!

Hi GB team, I could not find the iOS native sdk but only a flutter one, any plan to add it? Thx

One more question, do you have slides to introduce GB? Or anyone else in this community has done that? What I want to achieve is to share GB with my stakeholders to get their buy-in. Any help would be really appreciated

To deploy GB in AWS, what mongodb service do you recommend? 1. mongodb as a pod in EKS? 2. mongodb atlas - full managed service 3. AWS documentdb, another managed service 4 other options. I’m curious to hear your thoughts

Is there any (easy) way to have an A/B*/C* test? Our use case is to show some people A, some people B, and some people neither, and measure which works best. Thanks!

Hi, all. Simple question (I hope): I get presented with a login screen after connecting to my local docker url. Now.. how to log in? I managed to run it succesfully a couple of months ago, but cannot remember how I logged in back then. I'm using a fresh git directory, then docker-compose up -d

About user permission of the admin ui, is it possible to scope user based on project? For example, if user A isn’t in project B, he isn't allowed to change toggle value in project B.

Do we have functionality to remind people deprecate feature toggles if they are over their life time? I didn’t see any in doc

Hey GrowthBook team, thanks for your work on a very useful product! A few questions: 1. Why is it necessary to archive experiment results when increasing percentage traffic? From the dashboard: “Changing the traffic percent or split will start a new phase of the experiment. All previously collected results data will be archived and it will start fresh from this point on.” 2. Could you advise on the “correct” time to interpret experiment results? We’ve had an experiment drop in and out of statistical significance, looking at “Chance to Beat Control” in the dashboard. I figure it’s not great to look every day and decide the experiment’s done on the day it shows >95% 🙂

Hi team, if we have a strong need to isolate feature toggle data between environments, meaning in each environment we have to have a mongodb instance (or clusters), how to handle that scenario? Probably we need a super admin UI on top of all GB instances.

Hi GrowthBook team, I had a query regarding the storage of pvt keys of BigQuery. Are they stored in mongo? and If so what algorithm is used for encrypting it?

Hi team, after playing around with Growthbook and researching other open-source solutions out there, I was hoping you guys could help me with this question. English's not my native language, so my apologies if my wording strikes the wrong tone. What do you perceive as `growthbook`'s strengths compared to, say, unleash ? It seems to me the latter is quite mature in terms of feature toggling, but not as focused on integrating with data analytics platform as growthbook. It also has quite a lot of permission + performance features built-in that

growthbook

will probably want to adopt in the long run (project-specific, SSO, proxy cache etc.), so I guess it wouldn't be unfair for one to think why not go with the project that offers more.

Hello GrowthBook team, I'm evaluating Growthbook for our Data Team and I'm liking what I'm seeing so far. I have a question though. What would be the recommended way to integrate Growthbook with our existing analytics tools? For example, we have a simple tool to manage a definition of metrics (basically just a simple service + front end + backed by git repo), and we would like to keep all metric definition in one place if possible. So the ideal workflow that I can think of is for our analysts to continue defining metrics definition there, and push the changes into Growthbook when it happens through APIs. Is this possible? Or do you have other recommendation?

Hi GB team 👋 I have a couple of questions: 1. Is it possible to split experiment results by more than one dimension? Currently we have to select one or another from the dropdown menu, right? That would be a very useful functionality for our analysts. 2. When creating Dimensions & Segments we can only query those with the

user_id

. Would it be possible to enable Dimensions & Segments also for

anonymous_id

? Thank you very much!

Hello guys, I was looking around the ruby-sdk gem i made one pr, is any pr welcome ?

Hi guys~ Is there a helm chart to setup GrowthBook? currently we are planning to move it to EKS.

Hi Team , I am trying to setup growthbook feature in golang. I end up getting the default value irrespective of the split set. Any leads on how to fix this?

b, _ := json.Marshal(jsonMap["features"])
features := growthbook.ParseFeatureMap(b)
contextGB := growthbook.NewContext().
	WithFeatures(features).
	WithAttributes(growthbook.Attributes{
                "Key": ID,
            })
gb := growthbook.New(contextGB)
if gb.Feature("backend_enable_digilocker").On {
            log.Println("FEATURE IS ON")
}

Hi guys~ I noticed this SRM warning. but I’m not quite understand why it warns out. I feel the actual split show not big difference than expected.

Hi team, I'm unable to use "Force rules" and I think I'm doing something stupid 😅👇

Hello, there is any Tutorial to setup rails app to work with growthbook?

Hi team. we met an error when evaluating features. Any suggestions to bypass this error?

hey team! I was wondering if it’s possible to set up a test with more than 2 variations? (ie. I have a single variable that has 5 different states) Would want to see all 5 variations shown in an experiment at once. Is this possible?

Another question (and sorry if i’m spamming - new around here!) - is there a best practice for running a test where something other than a user determines the exposure? in my case, we are testing a variations of a widget, but a single user may see a different variation per session. i want to test the performance of each variation (based on some conversion metric). My initial thinking is that I can instead make the session_id be the user_id: so each session is bucketed instead of users.

Hi team~ found another UI bug. we have 5 projects. but the UI only shows 2 of them with the All. not sure how it happens. And when I switched to the

Projects

page. then the list back to normal. sounds like a cache problem.

Hello, just to be sure. Does the API always return

features

field in this case: https://docs.growthbook.io/app/api#feature-definitions-endpoint So, just to be sure that it is reasonable to expect that even if the

status

is something else than

200

there would be empty

{}

for features?

Hi. this might already have been asked so I'd appreciate a link if that is the case. Is there a way to define Features and Experiments in code (like a yaml or even API) and sync that to the Server to be reflected in the UI? .. I am thinking of like having those yamls versioned in github.. This sync could then be triggered on deploy of my new code (creating the new experiments) ,etc

In case you missed it, we released version 1.3.0 a few days ago. It has some cool improvements: • Customizable randomization units/Identifier Types • Feature versioning with audit logs and approval step • Multiple experiment assignment queries • Flutter SDK (contributed by @glamorous-florist-39551 and @gifted-cat-24658 from Alippo) • Sortable goal metrics If you're on our cloud version, you already have this latest version. If you're self hosting, be sure to update: https://docs.growthbook.io/self-host#updating-to-latest Please let us know if you have any questions or feedback.