I am having trouble with my Elasticsearch nodes/queries ……. I have 3 physical nodes in which all 3 are master + data nodes. I am running a simple filter query in slow logs I see query +fetch time as 8 millisec sometimes 17 millisec but response time on kibana is 300 to 500 millisec.
My index size is < 1 GB with 1 shard 2 replica, and the cluster has 16*3 GB RAM. Data size returned after filter query is around 1Mb(without whitespace).
Here is my slow logs
[2022-03-25T22:59:42,446][TRACE][index.search.slowlog.fetch] [debbf06e052fb476f254ba1234123] [khoj_prod_001][0] took[17.7ms], took_millis[17], total_hits[196 hits], types[_doc], stats, search_type[QUERY_THEN_FETCH], total_shards[1],
source[{
"size": 300,
"query": {
"bool": {
"filter": [
{
"term": {
"hub_id": {
"value": 1,
"boost": 1
}
}
},
{
"term": {
"category_id": {
"value": 3,
"boost": 1
}
}
},
{
"nested": {
"query": {
"bool": {
"filter": [
{
"term": {
"kml_wise_slot.kml_id": {
"value": "610fe6b10dab1241283de493",
"boost": 1
}
}
}
],
"adjust_pure_negative": true,
"boost": 1
}
},
"path": "kml_wise_slot",
"ignore_unmapped": false,
"score_mode": "none",
"boost": 1
}
}
],
"adjust_pure_negative": true,
"boost": 1
}
},
"sort": [
{
"scores.value": {
"order": "desc",
"nested": {
"path": "scores",
"filter": {
"term": {
"scores.key": {
"value": "ranking_scores",
"boost": 1
}
}
}
}
}
}
]
}], id[],