I'm having an issue getting the SDK set up in a Nest JS environment, but really all I'm trying to do is create a class that contains the configuration for growthbook and provides some wrapping functions. No errors are being thrown when initializing growthbook, but it always returns the default value passed to

getFeatureValue

- anyone else experienced this?

Hi everyone I have been facing an issue in js sdk and reaching out for help Issue: On updating a feature flag value, getting those values after couple of app session Details: I'm using the feature flag feature of GB in a react native application Currently I don't want to auto refresh the features in a single app session, so while initialising the sdk, I'm passing auto refresh as false and because of that no subscribedInstances are created. The issue is that when I fetch feature,

onNewFeatureData

is called and it sets the newly fetched features only when you have any

subscribedInstances

. Since I don't have any subscribed instances, the values are not set in the same app session but only updates the cache.

function onNewFeatureData(
  key: string,
  cacheKey: string,
  data: FeatureApiResponse
): void {
  ......
  updatePersistentCache();

  // Update features for all subscribed GrowthBook instances
  const instances = subscribedInstances.get(key);
  instances && instances.forEach((instance) => refreshInstance(instance, data));
}

Proposed solution: a straight forward solution would be to add a condition to check if auto refresh is enabled or not

if (instances) {
		//Update features for all subscribed GrowthBook instances
	  instances.forEach((instance) => setFeaturesOnInstance(instance, data));
	} else {
	  instance && data && setFeaturesOnInstance(instance, data);
	}

I wanted the understand the implications of these changes and is there any other way to solve this issue TIA!

Hi all. I have a question about

anonymous_id

. How to use it correctly? • For example, if I have an

orders

table and I build metrics based on it. • Do I need to store both

anonymous_id

and

user_id

in

orders

and the same fields in

viewed_experiments

? • Then add

anonymous_id

as Identifier in GB and select

ananoymous_id

from

orders

table in my sql metric? • And if the

anonymous_id

from

orders

and

viewed_experiments

match, then I will get a conversion, right? • What if a user from the orders table was previously assigned to a different option? For example, if I have this records in my database:

ORDERS TABLE (id, user_id, anonymous_id, timestamp)
1, 1, anon1, 2023-01-01 (the user id was known at the time of creation)
2, 1, anon2, 2023-01-04 (user ID added to the order after user authorization)

VIEWED_EXPERIMENTS TABLE (user_id, anonymous_id, experiment_id, variation_id, timestamp)
1, anon1, experiment1, 0, 2023-01-01 (the user created the order authorized)
null, anon2, experiment1, 1, 2023-01-04 (the user created an order as an unauthorized user and ended up in another group)

What is correct way to analyze experiment in such situation?

Any thoughts on how you would run cluster randomized experiments in GrowthBook?

Hello team, any insights about the

"Precondition failed"

error in my experiment results? Would be really helpful, thank you in advance.

how does growthbook compute over different datasources?

Hi all, I want to deploy growthbook on a server, and I have modified the docker-compose.yml, both the API_ORIGIN and API_HOST，but it failed. Anyone can help me?

Hey, I am trying to make some changes in the java sdk and test them. How have you created the jar from the code bcoz I seem to run into a lot of ClassNotFound errors for OkHttp and Gson transitive dependencies?

Hi, I'm hoping to get more information on how Sample Ratio Mismatches are detected. We have experiments that are flagged as having an SRM but the displayed splits are rounded to a 10th of a percent whereas the true splits should be irrational numbers (6 variants with equal allocation of the full audience) so I'm wondering if the SRM flag has something to do with rounding that is happening before the proportions test. Any info is much appreciated! (cc @Jane Carlen)

Hi GB team! We're close to getting A/B testing fully implemented. Currently implementing GA4 with a Firebase project but thinking if there's a quicker way without the scary thought of moving away from UA. Am I right in thinking you could have a function on your app push data into a db like BigQuery with just the barebones data needed since we already generate userIds. This would have the columns; userId, timestamp, experimentId, variantId and and event (the ones you want to track) and create the metric in the the dashboard from this data?