Test the performance of cbdatarecovery on magma buckets

The recovery of magma buckets has been enabled by this change: https://issues.couchbase.com/browse/MB-49475.

We should test the memory and CPU usage when recovering a magma bucket.

I have already tested this on my local machine with with a magma bucket that contains 1 shard and another one that contains 8 shards. The tests results are attached to this ticket with the script used to gather that data. NOTE: the script `mem_count_local_mac.bash` has been edited to work specifically on MacOS if on linux use `mem_count_linux.bash`

Further testing needs to be done with clusters that have around 100GB of data on them. The testing steps should be as follows:

1. Spin up a couchbase server on an AWS instance.
2. Create a magma bucket and using `cbpillowfight` generate 100GB of data on that bucket.
3. Run the script for collecting memory and CPU data (attached) and then run `cbdatarecovery`.
4. Collect the data.

We should do this for a magma bucket with 1 shard and 8 shards. We should then compare the results with the local run. We should determine whether the amount of data on magma affects the performance of `cbdatarecovery` and if it does look for ways to avoid/improve this.