Rationale
As seen below, we can deduce that the memory consumption of the reducers related to user status was the maximum, given that the app is idle with 50,000 users logged in and posting 35 messages per second. The memory consumption of the affected reducer (35%) is even more than that of the functions responsible for the incoming messages (27%). This implementation iterated over the received statuses multiple times across different reducers, each time processing a subset of the same data. This comes with the overhead of array function calls, creation of intermediary objects, and arrays.
Changes
Modified the getStatusesByIds action to process received statuses once, extracting and transforming all necessary data for statuses, dndEndTimes, isManualStatuses, and lastActivity reducer in a single iteration.
Improvements
By avoiding multiple iterations, we reduce the runtime complexity from O(4n) to O(n), where n is number of statuses.
Simplified reducer for user statuses, which includes statuses, dndEntTimes, isManualStatuses, lastActivity. Also created for single and multiple items of these reducers.
54d7011aba