Hi Everyone! We want to figure out how to disable dev tools growthbook extension/plugin so that in prod users can't see features we have + turn on/off. We set the following

window.growthbook_config.enableDevMode = false;
    window.growthbook_config.disableDevTools = true;

Just before our script

<script async
    data-client-key="<%= process.env.GROWTHBOOK_CLIENT_KEY %>"
    src="<https://cdn.jsdelivr.net/npm/@growthbook/growthbook/dist/bundles/auto.min.js>"
  ></script>

but it doesn't seem to do anything, still can access the dev tools extension, view & change values. Any ideas what we do wrong ? Thank you for help! 🙂

Hi All, I'm trying to create an experiment via the API POST request and have been running into some issues. Does anyone have a list of the minimum viable request body I can send to create an experiment? (I want to create a shell of an experiment just to try it out). Based on the docs I have an Experiment request containing the following fields with everything else being the default values shown in the example payload:

assignmentQueryId="user_id",
        trackingKey="example_experiment_trackingKey",
        name="example-experiment",
        project="my-project-id",
        variations=[variation1, variation2],
        owner="test-user@test.com"

However, I am now getting a message when I submit the request of the following:

{"message":"Unknown metric: string"}

since my default metric value is just "string". Based on the API Docs it doesn't looks like metrics is a required value. However, when I set

metrics = None

and pass that in the request body I get

{"message":"Request body: [metrics] Expected array, received null"}

Is there a default value I can pass for metrics, (and other), fields that will let me create a minimum experiment and is there a list of all fields that need to have valid data passed in as it seems that some fields which are not marked as required in the docs do require valid input. Thank you! If there is a better question for a different channel I am happy to post there, just let me know.

Hello GrowthBook Team, We are currently trying to integrate GrowthBook (self-hosted) into our service, but we’re facing an issue with the SDK. When we initialize GrowthBook in our repo, the feature flag evaluation always returns

false

, and the feature list is empty. However, if we check the API directly from the GrowthBook dashboard, we can see that the features are available, and the feature flag

FF

should return

true

. Here’s the snippet of our code:

routes.get('/test-ff', authMiddleware.closedAuth, async (req: any, res) => {
        const gb = new GrowthBook({
            apiHost: process.env.GROWTHBOOK_API_HOST,
            clientKey: process.env.GROWTHBOOK_CLIENT_KEY,
            enableDevMode: true,
            subscribeToChanges: true,
            attributes: {
                id: '123',
                country: 'US',
            },
            trackingCallback: (experiment: any, result: any) => {
                console.log('Experiment Viewed', {
                    experimentId: experiment.key,
                    variationId: result.key,
                });
            },
        });

        try {
            await gb.init();
        } catch (error) {
            console.log('Error', error);
        }

        if (gb.isOn('FF')) {
            res.send('Feature is enabled!');
        } else {
            res.send('Feature is disabled');
        }
    });

Do we need additional configuration for self-hosted GrowthBook in order for the SDK to fetch the features? Or is there something we might have missed here response from api (SDK Connection) on dashboard

{
  "status": 200,
  "features": {
    "FF": {
      "defaultValue": true
    }
  },
  "experiments": [],
  "dateUpdated": "2025-10-01T07:14:23.643Z"
}

Hello, i get this message on one of my experiment :

Multiple Exposures Warning. 114 users (14,82 %) saw multiple variations and were automatically removed from results. Check for bugs in your implementation, event tracking, or data pipeline.

My question is: How can GrowthBook determined that my users are seeing two versions of my AB-test ?

Hello GrowthBook! Can you help me to understand what am I need to do, to see some metrics in your dashboard? My app uses Next.js. I integrated your service on front-end part via "@growthbook/growthbook-react" Created experiment in dashbord. Created DataBase in BigQuery. In 'Data Sources' connected to BigQuery -> GA4. But when I tried to 'Run Analysis' in 'Metrics' I have error: "Error Processing Query Results Failed to run a majority of the database queries" What should I do? Am I need to create tables in DB by myself?

Hi All! Is the monthly pro subscription cancellable at any time? Is there a difference with pro per user pricing and licensing for self hosted? Thank you in advance

Happy Wednesday! We're switching sources for our testing data (from a BigQuery db to a RedShift db). My boss really likes seeing the historic status of the metrics and how they progress over the duration of the test. Will that be reset once I switch the data source?

Hello Growthbook Team 👋 I have some questions about the Hybrid (SSR + Client-side) approach you're presenting in your documentation. I'll leave them here: 1. Doesn't the

trackingCallback

get called two times? First, when the server prerenders the page, and then on the client side? 2. How is the server supposed to set the

attributes

when it initialize the Growthbook instance inside the

MyPage

component? I mean, in the example you clearly show that you set the

id

property by reading it from the cookies (i.e.,

Cookies.get('my_uuid')

), but when the server prerenders the page doesn't have access to cookies, at least inside the render function. Doesn't this involve a mismatching attribution between client and server, causing hydration errors? Thanks in advance 🙏

Is it possible to migrate / make sure that new Feature flags are created with 'v2' of hashing algorithm?

Hi all, not sure if this is a bug, but sometimes when I click "update" in an experiment the data does not actually refresh. I have to go to "re-run all queries." This is a bit of a pain because I use the API to update them all at once and pull new results into a spreadsheet. Now I have to go through and refresh all queries one at a time. Any advice?

Hey everyone, I’m having an issue with Growthbook and account access related to SSO. Here’s the situation: • I have a Growthbook account connected via Google SSO that gives me access to two organizations: Temple and Groover. • However, this SSO account is linked to the wrong Groover organization - one that has no activity. • I also have another Growthbook account (created with email + password) that’s linked to the correct Groover organization with all our data and activity. The problem is: when I try to log into that account, Growthbook forces me to use Google SSO, which redirects me back to the wrong organization. So I’m stuck in a loop and can’t access the correct Groover organization anymore. The only workaround I found is asking another admin to send me an invite again, but when I click on the invite link, I can only create an account with a password - and then I’m back in the same loop 😅 👉 Any idea how I could unlink the SSO or recover access to the correct Groover organization? Thanks a lot in advance! 🙏

Hi 👋 I’m not able to update my metrics, it’s running forever (I cut it off after 30 mins, right now it’s retrying again): https://growthbook.groover.co/experiment/exp_19g614mgf6k05r#. The test has only been live for a couple hours so it shouldn’t be much data. Any idea what’s wrong? Thanks!

Hello 👋 For context, we recently deployed a new search page for our hotel booking app using GrowthBook. I ran an experiment and added two metrics — hotel click and checkout — which are based on GA4 events. However, when I check the results in GrowthBook, I’m not seeing any data. But when I use the SQL Explorer, I can see the data there. So I’m sure I’m missing something in how GrowthBook handles metrics. Could you please guide me on what I might be doing wrong? Thanks!

Hi! 👋 I've noticed that Growthbook often logs out unexpectedly during work - you might be configuring a test, and suddenly the session ends, losing all your changes. If you reload the page, it immediately logs you out as well. This happens quite often now: while working on tasks, we have to log in again and again - sometimes up to 10 times a day. It makes testing harder and slows down the workflow, especially when changes are lost. Has anyone else faced this issue or maybe has an idea how we could fix it? 🙏🏻

Does anyone know the experiment targeting rules for user attributes that change value? e.g. • If a user doesn't initially match the target attributes - the first time they would otherwise be in the experiment - can or should they be included in the experiment if they match later on? • Once a user has matched attributes and been included, do they continue to see experiment variants if the targeting conditions subsequently fail? I can see in general it's a lot of complexity having changing user attributes although in some cases it could be quite useful

Hi all! Im trying to set up gb with php, using a simple bool feature flag. Everything is working in dev but in production I don't see any changes on the page, in the growthbook extension I can see that the flag is true. Don't know how many php-users are here, and I'm sitting in a project which is a bit dated in places, but I'm using pdoCache wrapped in Psr16Cache to set the cache, and using the initialize method to load features. Anyone have any thoughts on what I might need to do to get it to work in prod? When I use the Growthbook extension I also cant see my backend SDK there, but that happens in dev as well (even though the test is working fine ther)

I am unable to see results for the experiment. nor i am seeing any experiment in debug mode in GA4. Is there any person who can help me with this from the team ?

We’re running redirect-based testing, but I’ve noticed that our Facebook pixels (and other tags) are firing on both pages. What’s the best practice to prevent that from happening? We don’t have full control since we’re using a landing page builder, so going fully server-side isn’t an option...

Hi GB team, we are trying to get sticky bucket working on both ios and android SDKs. We have previously identified issues with race conditions and have worked with you all to get those issues addressed. There are now PRs open on both ios and android (kotlin here, swift here) that fix the issues we saw. Just wondering now how long it might take to get these PRs tested, merged and released? We are eagerly awaiting closing the book on our sticky bucketing implementation and using it in some experiments that are currently on hold until the fixes are released 🙂 thanks for the help to date!

Hi all, I am trying Self Hosting GrowthBook with kotlin backend, initialize gb sdk, but failed to get feature value

image.png

and failed to connect sdk

Hi guys My Saved Groups (ID Lists, id key) with Feature Flags suddenly stopped working. Any changes on your side?

Are the columns available in the custom SQL filter on the analysis settings of an experiment strictly limited to those listed?

Hi there, We are evaluating several feature management and experimentation platforms for a key business initiative, and your solution is on our shortlist. We are looking for a server-side tool that allows our team to manage URL redirects from a simple dashboard. Our primary use case is to route traffic from a single starting URL to different destination pages based on URL parameters. The unique challenge is that the tool's SDK will be implemented on a third-party partner's website, requiring a clear separation between their technical setup and our ongoing campaign management. Our use case is very specific and has some critical technical requirements. Before we proceed, we need to confirm if your platform can support our exact needs. Our Operational Model & Core Requirements: 1. Partner Implementation & Support: We will be working with a large partner company whose developers will perform a one-time, server-side integration of your SDK into their blog's backend. ◦ Onboarding: Do you offer free onboarding support to guide our partner's development team through this process? What are your typical support response times for implementation-related questions? ◦ Setup Time: What is the average time required for a developer to complete this initial SDK integration? ◦ Security Scoping: Is it possible for the developers to limit the SDK's execution to a specific URL path (e.g., only run on

<http://partner.com/blog/*|partner.com/blog/*>

)? This is a critical security requirement to ensure our rules can only affect a designated part of their website. 2. Separation of Concerns (Critical): Our operational model requires a clear separation of duties. The partner's developers are responsible for the one-time technical integration, while our team manages the ongoing campaign strategy. Therefore, the platform must ensure that our campaign rules, targeting logic, and destination URLs are managed exclusively through your dashboard and are not visible or configurable within the partner's backend code. 3. Server-Side Execution: The entire process must be server-side to avoid any client-side "flicker effect." Client-side JavaScript-based redirects are not a viable solution for us. 4. Primary Use Case (Rule-Based Redirect): We need to route traffic from a single starting URL to various destination landing pages based on URL query parameters. For example: ◦ A user visiting

<http://partner.com/base-url|partner.com/base-url>

(with no specific parameters) sees the default landing page. ◦ A user visiting

<http://partner.com/base-url?utm_campaign=campaign_1|partner.com/base-url?utm_campaign=campaign_1>

must be instantly redirected to

<http://partner.com/lp1|partner.com/lp1>

. ◦ A user visiting

<http://partner.com/base-url?utm_campaign=campaign_2|partner.com/base-url?utm_campaign=campaign_2>

must be instantly redirected to

<http://partner.com/lp2|partner.com/lp2>

. 5. Parameter Stripping & Replacement (Critical): When a redirect occurs, we must have full control over the final URL's query parameters. The original trigger parameter (

?utm_source=meta

) must be completely stripped, so the end client cannot see our operational logic in their analytics. We must be able to define the final URL to be either a clean URL (

<http://client.com/product|client.com/product>

) or one with custom-set parameters (

<http://client.com/product?utm_source=partnerblog|client.com/product?utm_source=partnerblog>

). Our Questions: 1. Can your platform fully support the operational model and all of the core functional requirements listed above? 2. Regarding Requirement #5 (Parameter Stripping), does your platform offer a native feature to manage this server-side, or would this require custom logic to be written by the implementing developers? 3. What are your typical support response times for ongoing technical inquiries (post-implementation) for a customer on your free plan? This is a key strategic project for us, and the platform we choose will be a long-term partner. We appreciate you taking the time to address these specific points. Best regards, Moritz

Hi all, we are getting our GrowthBook setup with our first experiment. We have managed to track the proportion of visitors that create a lead on our site, so we can see the increase in the number of users who create a lead through a Metric setup as a Proportion. What I cannot figure out is how to do a simple count of the number of leads. This is important to us because a user might create multiple leads in the same session or over many sessions.

Hi, We’re running into an issue under the Results tab in the UI. We see this error:

Error=2, Details='Response status code does not indicate success: BadRequest (400); ... {"Errors":["The index path corresponding to the specified order-by item is excluded."]} ...'

We’re using a MongoDB instance in Cosmos with these settings:

"EnabledApiTypes": "MongoDB",
"apiProperties": { "serverVersion": "7.0" }

Any idea what might be causing this?

If someone finds our growthbook's api key (for the connection) and tries to abuse, and we delete the old keys, if they request using the old keys even though they're deleted, do we still get billed for that? I'm seeing that we're billed, but that doesn't make sense, is it a bug in GrowthBook?

Hi, I have delta tables in my org, can someone please help me or direct me on How to add these delta tables onto Fact tables or Data sources in Growth Book. I am Stuck here. Thanks in advance!

When I force a version. Either via ?forcevariation= in the URL or via match in a UTM via the rules on the feature in growthbook. Should I expect the trackinig event to send again and with the new version the user is put in via one of the forced rules?