The concept being tested is how to cap parallel work inside a Scatter-Gather. Scatter-Gather splits a message into multiple branches and processes those branches in parallel. To prevent overwhelming downstream systems or the Mule runtime, you can limit how many branches run at the same time. The maxConcurrency property provides that precise control: it sets the maximum number of concurrent branch executions. If there are more branches than this limit, the extras wait in a queue until a running branch finishes. Other terms don’t directly apply as the per-scatter limiter. Concurrency is a general idea of parallel work, but the specific property used to bound Scatter-Gather execution is maxConcurrency. ThreadPoolSize relates to the underlying thread pool used by the runtime more broadly, not the per-scatter concurrency limit, and parallelism isn’t the property used to gate the scatter’s parallel processing.

Object Store differences come down to capacity and runtime compatibility. Anypoint Object Store V2 is designed for Mule 4 and offers larger limits for stored data, making it a more scalable option for modern flows. It also operates specifically with Mule 4, so it isn’t intended for Mule 3 deployments. CloudHub Object Store V1 is the older, Mule 3-era store with smaller limits, used in CloudHub environments. It isn’t the newer, higher-capacity option and isn’t the Mule 4-focused store. Additionally, these stores are cloud-managed services, not on-premises solutions. So the main distinction is that Object Store V2 provides larger limits and is Mule 4 only.

DataWeave is used to transform and shape data in Mule flows. When dealing with paginated responses, the goal is to present a consistent, client-friendly structure regardless of how the backend returns data. DataWeave takes the backend payload—items plus pagination metadata—and maps it into a standard paginated envelope, such as including the data list, total count, current page, page size, and total pages, often with navigation links. This transformation ensures clients always see a uniform paginated response, which simplifies consumption and keeps behavior predictable across different backends. The other tasks—storing user sessions, rendering HTML for API docs, or handling security tokens—lie outside this transformation role and are addressed by other parts of the MuleSoft stack.

The two main phases are: first, the Authorization Code Flow, where the user authenticates and an authorization code is obtained via the authorization server; second, the Exchange Authorization Code for Access Token Flow, where that code is sent to the token endpoint to obtain an access token (and optionally a refresh token). This separation keeps user credentials and tokens secure by using a back-channel exchange rather than exposing tokens to the user-agent. Other flows mentioned correspond to different OAuth 2.0 patterns and do not describe these two phases of the Authorization Code grant.