When memory usage estimation API is called with model_memory_limit set to a low value, the following error is returned by the process:

Input error: memory limit 0.89MB is too low to perform analysis. You need to give the process at least 4MB, but preferably more.

This failure is a combination of 2 factors:

1. computeAndSaveExecutionStrategy method that validates the model memory limit condition is invoked as part of CDataFrameAnalysisRunner initialization:
   https://github.com/elastic/ml-cpp/blob/master/lib/api/CDataFrameAnalysisRunner.cc#L201
2. the model memory limit in memory usage estimation request is always set to 1MB

There are 2 solutions I see:
[simple] send large model memory limit in memory usage estimation request so that the validation passes
[more involved] do not call computeAndSaveExecutionStrategy method at all in memory usage estimation process