I am still using Elasticsearch 5.6 version
I have an index with a completion suggest field and also a text field with edge ngram size 3 to 10 [i don't query both of them at the same time]…. so I got few queries which use the completion suggest field and few which use the edge ngram text field
lately I have been seeing memory usage/cpu spike on the nodes with shards from this index ... and these nodes are experiencing out of memory/garbage collection issues.. eventually crashing.
So clearly we are having issues with this index… do y'all think its because I have both completion suggest field and edge ngram field on the same index? [i know it sounds silly] or is it it one of these fields is causing the issue
more info: this index is like 220mb, 56k records with 3 shards one replica… also like 20k requests per hour
yesterday i updated this to only have one shard instead of three, to see if that helps.
any insight is greatly appreciated! thanks!