Best practices for integrating Split feature flags into a serverless environment

In serverless environments, data persistence is best handled by externalizing state. This avoids the performance hit of "cold starts" where processes have to load and cache data before they can perform. This is the case with feature flagging SDKs as well.

Read the blog post, Serverless Applications Powered by Split Feature Flags for examples of how to achieve the best performance when using Split from within Lambda functions in Amazon AWS.