Versions (relevant - OpenSearch/Dashboard/Server OS/Browser):
OpenSearch 1.1
Describe the issue:
Intermittent slowness on the query response.
Configuration:
Cluster :
data nodes
Instance type c5.large.search (3)
Dedicated Master Node : r4.large.search (3)
Storage Type : EBS
EBS Volume Type : General Purpose (SSD) - gp2
EBS Size : 10 GiB
Relevant Logs or Screenshots:
Collection size is 1MB ; with below settings:
{
“spec_proc_comb_exp” : {
“settings” : {
“index” : {
“refresh_interval” : “86400s”,
“number_of_shards” : “5”,
“plugins” : {
“index_state_management” : {
“rollover_skip” : “true”
}
},
“provided_name” : “spec_proc_comb_exp”,
“creation_date” : “1671629334835”,
“number_of_replicas” : “2”,
“uuid” : “aht7O4QQTV6WtozcVCfi1A”,
“version” : {
“created” : “135227827”
}
}
}
}
}
Query run:
GET spec_proc_comb_exp/_search
{
“query”: {
“bool”: {
“must”: [
{“multi_match”: {
“query”: “dent”,
“fields”: [“name”,“alias_terms”],
“fuzziness”: “4”
}}
],
“filter”: {“match_phrase”: {
“category”: “Specialty”
}}
}
}
}
Problem:
We are using OpenSearch as backend to perform exact/fuzzy match for a UI search bar. The index currently used is pretty small 1MB. We see an issue with not stable Response time. Most of the time it is around 100ms; but at times intermittently (10% ) is around 1000 ms or higher.
Request experts to please help troubleshoot this issue. I am new to OpenSearch or any search tech.