Closed
Description
It is useful to know the total amount of native memory ML is allowed to use cluster-wide at a given time, for example when planning which subset of a large number of security jobs to enabled.
It is possible to calculate the quantity in a client given knowledge of the node attributes and cluster settings. However, this breaches encapsulation by duplicating a formula that is used in Elasticsearch code, and runs the risk that client and server diverge in the future.
Therefore, it would be better if the result of the calculation done in Elasticsearch code was made available as one of the outputs of the ML info endpoint.