Hi everyone,
We are currently running on an Enterprise ERU (capacity-based) license IN SELF HOSTED ON PREMISES CLUSTERSl.
Based on my understanding, Master, Data, and Machine Learning (ML) nodes are the primary nodes that count toward consumption of ERU and as per the Standalone license model - #2 by stephenb .
However, I want to confirm if the following dedicated utility nodes are excluded from counting against our ERU license pool, or if they consume it based on their exact allocated RAM:
- Ingest & Coordinating Nodes with remote_cluster roles as ccs enabled for clusters
- Transform Nodes with with remote_cluster.
- The ML Node Query: For Machine Learning nodes, the native ML jobs execute as C++ processes outside of the JVM on our infrastructure, requiring only a tiny JVM heap (around 2GB) for cluster coordination. Under the ERU license model, is our ERU consumption calculated strictly based on the Elasticsearch JVM Heap size assigned to the ML node, or is it calculated based on the total physical memory/container RAM allocated to the machine hosting the native C++ processes?
We want to safely increase our ingest, coordinating, and transform layers to protect our data nodes, but we need to know exactly how Elastic's license accounting tracks these specific roles under the ERU model.
Thanks!