Hi All,
I have a production cluster with 4 nodes each have 24 GB RAM and heap is 12GB on for each node.recently i have configure circuit breaker to to avoid OOM exceptions because of heavy field data loading. i have use the following properties .
fielddata
indices.fielddata.cache.size: 5gb
circuit breaker to limit field cache to avoid OOM exceptions
indices.breaker.fielddata.limit: 7gb
indices.breaker.request.limit: 5gb
indices.breaker.total.limit: 8gb
so i have set limits of circuit breaker like above. let say if am getting query exceptions thrown by circuit breaker with above configurations, if a field is taking more memory to process aggregations etc...in this case what we have to do? do we have to increase our current memory of each node or we need to scale out cluster by adding one more node with same configuration? If we need to scale cluster How elastic search will balance this field data cache memory load across cluster?.please explain me this scenario.
Thanks & Regards,
Phani