elasticsearch 调优
作者：互联网
elasticsearch 调优
1、 内存占比很高，硬盘占比较高，降低内存和硬盘占比情况
2、 在elastic search application.yml增加配置，关闭 Swap
bootstrap.mlockall: true
3、数据压缩传输配置
http.compression: true
http.compression_level: 1
4、Nlp 和Elasticsearch 中的http使用连接池和keep alive
5、不需要传的字段不要传，尽量从es获取数据，不要从多个接口获取数据
6、修改刷新频率
PUT michaels_product/_settings
{
“index.refresh_interval” : “30s”
}
7、内存配置要机器内存的一半
8、避免日志数据无限的增长，只保留5天日志，zplin的数据信息独立分开到别的服务器
9、降低内存使用的占比，最好要在75%以内，避免响应时间慢
10、读写分离、一个节点读，另一个节点写
11、关闭大量的轮询更新，部分轮询更新改为pub/sub增量更新
12、对索引进行translog的优化。
PUT /michaels_projects/_settings
{
“index” : {
“refresh_interval” : “60s”,
“translog” : {
“generation_threshold_size” : “64mb”,
“flush_threshold_size” : “1024mb”,
“sync_interval” : “60s”,
“durability” : “async”
},
“merge” : {
“scheduler” : {
“max_thread_count” : “1”
},
“policy” : {
“segments_per_tier” : “30”
}
}
}
}
PUT /michaels_events/_settings
{
“index” : {
“refresh_interval” : “60s”,
“translog” : {
“generation_threshold_size” : “64mb”,
“flush_threshold_size” : “1024mb”,
“sync_interval” : “60s”,
“durability” : “async”
},
“merge” : {
“scheduler” : {
“max_thread_count” : “1”
},
“policy” : {
“segments_per_tier” : “30”
}
}
}
}
PUT /michaels_product/_settings
{
“index” : {
“refresh_interval” : “60s”,
“translog” : {
“generation_threshold_size” : “64mb”,
“flush_threshold_size” : “1024mb”,
“sync_interval” : “60s”,
“durability” : “async”
},
“merge” : {
“scheduler” : {
“max_thread_count” : “1”
},
“policy” : {
“segments_per_tier” : “30”
}
}
}
}
PUT /michaels_browsing_history/_settings
{
“index” : {
“refresh_interval” : “60s”,
“translog” : {
“generation_threshold_size” : “64mb”,
“flush_threshold_size” : “1024mb”,
“sync_interval” : “60s”,
“durability” : “async”
},
“merge” : {
“scheduler” : {
“max_thread_count” : “1”
},
“policy” : {
“segments_per_tier” : “30”
}
}
}
}
PUT /michaels_category/_settings
{
“index” : {
“refresh_interval” : “60s”,
“translog” : {
“generation_threshold_size” : “64mb”,
“flush_threshold_size” : “1024mb”,
“sync_interval” : “60s”,
“durability” : “async”
},
“merge” : {
“scheduler” : {
“max_thread_count” : “1”
},
“policy” : {
“segments_per_tier” : “30”
}
}
}
}
PUT /michaels_promotion/_settings
{
“index” : {
“refresh_interval” : “60s”,
“translog” : {
“generation_threshold_size” : “64mb”,
“flush_threshold_size” : “1024mb”,
“sync_interval” : “60s”,
“durability” : “async”
},
“merge” : {
“scheduler” : {
“max_thread_count” : “1”
},
“policy” : {
“segments_per_tier” : “30”
}
}
}
}
PUT /marketplace_makers/_settings
{
“index” : {
“refresh_interval” : “60s”,
“translog” : {
“generation_threshold_size” : “64mb”,
“flush_threshold_size” : “1024mb”,
“sync_interval” : “60s”,
“durability” : “async”
},
“merge” : {
“scheduler” : {
“max_thread_count” : “1”
},
“policy” : {
“segments_per_tier” : “30”
}
}
}
}
PUT /marketplace_product/_settings
{
“index” : {
“refresh_interval” : “60s”,
“translog” : {
“generation_threshold_size” : “64mb”,
“flush_threshold_size” : “1024mb”,
“sync_interval” : “60s”,
“durability” : “async”
},
“merge” : {
“scheduler” : {
“max_thread_count” : “1”
},
“policy” : {
“segments_per_tier” : “30”
}
}
}
}
PUT /product_autocompletion/_settings
{
“index” : {
“refresh_interval” : “60s”,
“translog” : {
“generation_threshold_size” : “64mb”,
“flush_threshold_size” : “1024mb”,
“sync_interval” : “60s”,
“durability” : “async”
},
“merge” : {
“scheduler” : {
“max_thread_count” : “1”
},
“policy” : {
“segments_per_tier” : “30”
}
}
}
}
PUT /order_history/_settings
{
“index” : {
“refresh_interval” : “60s”,
“translog” : {
“generation_threshold_size” : “64mb”,
“flush_threshold_size” : “1024mb”,
“sync_interval” : “60s”,
“durability” : “async”
},
“merge” : {
“scheduler” : {
“max_thread_count” : “1”
},
“policy” : {
“segments_per_tier” : “30”
}
}
}
}
PUT /return_order_history/_settings
{
“index” : {
“refresh_interval” : “60s”,
“translog” : {
“generation_threshold_size” : “64mb”,
“flush_threshold_size” : “1024mb”,
“sync_interval” : “60s”,
“durability” : “async”
},
“merge” : {
“scheduler” : {
“max_thread_count” : “1”
},
“policy” : {
“segments_per_tier” : “30”
}
}
}
}
PUT /event_autocompletion/_settings
{
“index” : {
“refresh_interval” : “60s”,
“translog” : {
“generation_threshold_size” : “64mb”,
“flush_threshold_size” : “1024mb”,
“sync_interval” : “60s”,
“durability” : “async”
},
“merge” : {
“scheduler” : {
“max_thread_count” : “1”
},
“policy” : {
“segments_per_tier” : “30”
}
}
}
}
PUT /dispute/_settings
{
“index” : {
“refresh_interval” : “60s”,
“translog” : {
“generation_threshold_size” : “64mb”,
“flush_threshold_size” : “1024mb”,
“sync_interval” : “60s”,
“durability” : “async”
},
“merge” : {
“scheduler” : {
“max_thread_count” : “1”
},
“policy” : {
“segments_per_tier” : “30”
}
}
}
}
PUT /order_item/_settings
{
“index” : {
“refresh_interval” : “60s”,
“translog” : {
“generation_threshold_size” : “64mb”,
“flush_threshold_size” : “1024mb”,
“sync_interval” : “60s”,
“durability” : “async”
},
“merge” : {
“scheduler” : {
“max_thread_count” : “1”
},
“policy” : {
“segments_per_tier” : “30”
}
}
}
}
标签：60s,settings,interval,refresh,调优,elasticsearch,threshold,size 来源： https://blog.csdn.net/laowaizhangge/article/details/122232388