hi everyone! is there a way to set different default values for the same flag for staging and production?

Hello ! Is it possible to get in contact with the "Java expert"? I am having trouble with the encryption for the webhook and I am pretty sure my issue can be solved by someone that had experience setting up webhooks with Java

Hi, we recently found that growthbook would throw fatal error in incognito mode. Because it tried to use localStorage

A/B test ramp up question. It seems that increasing the sampling % in the experiment underlying feature doesn't increase the experiment sampling % ? We need to adjust the experiment sampling % directly? Would there be any trouble if the underlying feature sampling % and Experiment sampling % are differnt ?

hey all, I don’t know who runs this Slack but I accidentally clicked a button to add an app, you all can ignore it (I thought it was asking for my company Slack) sorry for the noob mistake 😆

Hi all, quick questions. will GrowthBook SDK auto refresh bring too much pressure to the server if there are like more than 100k client using it? Because I see it’s listen on SSE events.

If I convert from a rollout rule to an experiment rule, will the groups of users be the same?

If I use the deviceId to split users in a feature, how would this work for events that are triggered in mixpanel with the distinct id for example?

Hi , how exactly does growthbook work with google analytics ga4 ? I am sending my variation id and experiment id in an event name called viewed_experiment , and I am having a separate purchase event which is being send to analytics. My question is how does growthbook indentify from which variant the purchase has happened?

It looks like, in order to remove a data source, you have to delete it. To do that, you have to delete all its metrics. But if you still want to preserve the historical experiments, it seems like you'd be stuck? Is it possible to remove the data source without breaking those? Or could it be an option to archive a data source instead?

Hi there. My experiment is showing “not enough data” even though I’ve reached the minimum sample size specified in the settings. Are there some other factors that come into play?

Are there any plans of adding a possibility of applying corrections to control family-wise error-rate in case of testing multiple outcomes? I feel like I may have trouble trying to communicate to the management that the chance to beat control of 95% when there are 3 metrics being tested is not too reliable

Hey there, I just tried integrating the trial license key into growthbook through settings as mentioned in the docs, but once I entered it’s just stuck on this now and not loading, can anyone please help?

Hi, is it possible to configure the frontend to call API endpoints behind

/api

instead of on a separate port or subdomain? Tried out:

APP_ORIGIN=<http://some.url.com>
API_HOST=<http://some.url.com/api/>

But the frontend seems to handle it inconsistently

Hi, also an API question, we have deployed gb in AWS ECS and want to use CloudFront to cache the API, our setup is:

APP_ORIGIN: <http://growthbook.ourdomain.com|growthbook.ourdomain.com>

API_HOST: <http://growthbook.ourdomain.com:3100|growthbook.ourdomain.com:3100>

and then we have the CloudFront distribution created as

<http://growthbook-api.ourdomain.com|growthbook-api.ourdomain.com>

The issue now is that we cannot enter

<http://growthbook-api.ourdomain.com|growthbook-api.ourdomain.com>

into the API_HOST environment variable. I'm guessing that is because of the port mapping from container port to host port (or loadbalancer target to container even). As we also can't use any other port than 443 for CloudFront, we're currently not sure how to solve this. Do you have a working example for the GrowthBook API behind CloudFront?

Hey here! Hope you are doing well. I have a question about the RAM consumption. I’ve deployed the

growthbook

application from the

growthbook/growthbook:1.9.0

image into the AWS ECS Cluster. During the approximately 10 days, consumption of the memory has been increasing (you can see it on the attached screenshot). Now it has been already reached 80% of RAM. It is without any traffic for the containers. For container I specified 512MB of memory, as soon as haven’t found in the official documentation, how many memory should I provide for container. So my question is, don’t you know why container consumes RAM more and more with each day? Maybe you faced with that? Thanks!

Hello team, if my experiment query has an error that says "Precondition failed" it's a duplicate of another experiment that doesn't have the error, what could be the problem? Thank you in advance.

We're using the Ruby SDK for Growthbook and we want to track/create experiments with a data source of Mixpanel. We've followed the documentation closely but we keep seeing the following error. Our VariationID is

0

or

1

- is this why we're seeing this error? Would be good to get an explanation on the below error message!

hello @future-teacher-7046 @fresh-football-47124 i am running an A/A split test - set the feature to 50/50 split -but it seems that when i am doing the stats the sampling for each variant is not even close to each other A1 - 317 A2 - 259 Although i dont have an SRM but question is is this normal?

or is there a way to make the sampling for each version closer to each other?

Hi, how do I put an auth behind :3100 port which serves the api information? as in our production deployment we don’t want it to be open publically, now say we are able to put auth, then how will growthbook dashboard still work with this?

Hi team, we are using big query data source. I found that when I add a boolean denominator to a metric, it would cause the experiment display no data. Also, when I added activation metric to the experiment, it would said no data too. Could anyone help on troubleshoot that?

Hi Team I have question. We ran an experiment and had a clear winner but some of our metrics have 7 day conversion window. We released the feature to 100% users within few hours of conclusion of the experiment. On day of conclusion the chance to beat control was 98% with 27% mean relative uplift. The same day the feature was released to 100% users including those who are part of the control group. We see that after release the gain is minimising We have 2 day conversion window and 7 day conversion window for same metrics. After 48 hours of the experiment conclusion - the mean relative uplift is 15.9% with probability to beat the control is 95.83% 1. Will it affect our metrics as the users from the control group are also released to the new the feature and some of them may still fall under the 7 day conversion window? 2. Can this be due to the users who are under the control group are now exposed to the new feature and are converting hence leading to the reduction in the gain that we had initially?

@fresh-football-47124 why do we use 95% credible interval and not 89% . I was reading in some papers for bayesian framework 89% CI would be a better compared to 95% though arbitrary?

Hi everyone, I've been using Growthbook feature-flagging, more as a way to configure my projects(a replacement for my .env files basically). For example, what index to search in, when a search request comes in, or what filters to show each customer. Do you think this use case is logical, or is there a better way to do this, and I'm just trying to use Growthbook for everything? I was thinking, because Growthbook provides an api, i can also build an interface that allows even non-technical people in the company to configure all our projects, without waiting on the developers.

Besides, there is one more issue that we faced. We deployed the app on one of our servers and it worked properly for a couple of weeks but after that it's giving timeout. We even tried starting it again but no luck still timeouts. Any suggestions please??? Thanks in Advance

Hi, i was setting up a project in the growth book, the feature split is 80,10,10 in the feature rule, but when I go to experiment there also I have the option to change the split so what is the relation between those?

Hi, hoping you all can help us understand a funny output of the allocation model. Seems the allocation model is assigning similar (but different) ids all to the same side. These ids are different by the last digit but don’t seem to be getting allocated evenly across the two sides. Any idea whats going on here.

We have hosted the app on one of our machines using docker. To access it via IPs we have added all the suitable env variable like(api_host,app_host,jwt etc.) It was working like a charm for more than a fortnight. Now suddenly since yesterday,app(both ui i.e. port 3000 and api 3100) are throwing connection timeout. when we try telnet(connection check) from other machine it is not connecting. On that (host) machine also if we try curl using localhost it gives a response which means service is up. But if we try the same using IP ,instead of localhost it doesnot return any response. Note :- No error in either of the container's logs. Please Help out if any suggestions. Thanks in advance. (edited)