Question 1

How does Flagsmith handle caching?

Accepted Answer

All evaluation is done on our servers. That way you don't leak segment definitions to the client.

In terms of caching, by design we do not have any caching infrastructure. You are free to proxy our API via a CDN if you wish. There is also a good pattern around caching flags within your infrastructure using our webhooks here:

https://docs.flagsmith.com/integration-approaches/#caching-flags-on-a-server

Question 2

Does the Open Source version of Flagsmith have any restrictions?

Accepted Answer

The entire codebase of Flagsmith is publicly available. The Flagsmith product has a dual license for use, with the majority of the codebase being Open Source under a BSD 3-Clause license.

The Enterprise Edition (EE) repository holds a portion of code for features we sell under an Enterprise license. These may not be used in production without an Enterprise license from Flagsmith. This includes RBAC, and other features we may include in this license in the future.

You can see which repositories are in which license on GitHub: https://github.com/Flagsmith

Question 3

Does Flagsmith support SSO?

Accepted Answer

Yes we support Google and Github SSO on all of our plans including the free tier.

Question 4

Is Flagsmith SOC 2 certified?

Accepted Answer

Flagsmith is SOC2 Type 2 certified. We can provide our SOC2 audit report on request.

If data security is a top priority we recommend using Flagsmith On-Premises. We support this for our Enterprise clients, and work with large banks and healthcare providers that have chosen us because this gives them the highest level of control and security.

Question 5

How do I identify an anonymous user? What about running AB tests with anonymous users?

Accepted Answer

In this case you should generate a GUID/hash for the user on your client side, cookie it, and use that to identify the user.

In terms of running tests against anonymous users, Flagsmith needs some sort of consistent ID that you store against the user on the client, otherwise it's impossible for us to serve a consistent test to that user. Generally people just create a random GUID and then store it on the client (e.g. using cookie on web) but it depends on your client implementation.

Question 6

Is there a way to control who gets a feature flag by percentage?

Accepted Answer

Yes. To do this you would create a Segment with a percentage split rule. From there you can set a Segment Override to enable or disable the flag for that percentage split segment.

More details on implementing percentage splits here: https://docs.flagsmith.com/staged-feature-rollouts/#creating-staged-rollouts

Frequently Read Articles

Browse All Categories