For machine learning, how many ML/ingest nodes are good for 10 data nodes (3 of them are master eligible nodes)?


I posted a couple questions here and still considering about applying platinum license for anomaly detection feature.

10 elasticsearch nodes will have the license and they will be used only for "anomaly detection" + "automated reporting" features.

I am planning to configure...

All 10 nodes as Data nodes, some of them set as ML nodes, and 3 of them will be configured as Master eligible nodes.
Coordinating + Ingest nodes will be added additionally.

Is there best guidance for my case about # of ingest nodes? How many ingest / ML nodes are good for set in case they need to work with 10 data nodes?

they can be set as multiple roles configuration. right? (e.g. node works as master/data/ml)

Please advise.

Thank you!

This blog ( has good information on best practices with sizing ML nodes with respect to the cluster size.

thank you for your advice. the link you provided was very helpful.

thank you!

1 Like

This topic was automatically closed 28 days after the last reply. New replies are no longer allowed.