Minumum ram size for vector model ( e5, elser)

Hello im trying to deploy ml model on my docker es and kibana
for Test.

and i dont know how much ram size is required for run this model as ingest pipeline .

i trying to deploy it with kibana on this site
: Machine Learning> Model Management > Trained Model

but it says

not enough memory on node [XD9jqzjXQriFeInG5UPRRA] to assign [1] allocations to deployment [.elser_model_2_with_4g]

my docker es machine has 4GB of heap size
this is result of /localhost:9200/_cat/nodes?h=heap*&v

/localhost:9200/_cat/nodes?h=heap*&v
``` 

heap.current heap.percent heap.max
512.1mb 12 4gb

Hi @dan_kim our documentation says that 4 GB is the minimum for ELSER, but that may depend on what else you're doing. Can you please share more about your current memory usage?

Also to clarify, the 4GB allocation is on your ML node?

Finally, is turning on autoscaling an option?

1 Like

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