WHY
Alert firing on GCP installs

WHAT
We currently calculate the free memory of a Postgres instance and use this to alert if the free memory is less than 10%.

The metric fires based on the following formula:

(cro_postgres_freeable_memory_average / (1024 * 1024)) < ((cro_postgres_max_memory / 100) * 10)

This assumes that the free memory is provided in bytes but the max memory is provided as mibibytes. The GCP implementation assumed that both were provided as bytes so we should standardise on a format.

HOW
Update the AWS implementation to use bytes by changing the value here in CRO

Update the alert to assume that the format of the memory values is now bytes for both.

cro_postgres_freeable_memory_average < ((cro_postgres_max_memory) * 0.10)

Also update this dashboard as it will no longer need to convert to bytes

TESTS
<List of related tests>

DONE
Both the alert and dashboard are accurately tracking the postgres memory using the same formats.