Elk Architecture: Elasticsearch+kibana+filebeat
Version information:
Elasticsearch 5.2.1
Kibana 5.2.1
Filebeat 6.0.0 (preview)
Today in the Elk Test, the Kibana above the discover regardless of the index, found that will be error:
[Request] Data too large, data for [<agg [2]>] would is larger than limit of
And in the Elasticsearch log you can see:
Org.elasticsearch.common.breaker.CircuitBreakingException: [Request] data too large, data for [<agg [2]>] would be Larger than limit of [1283260416/1.1GB]
According to the error message with Google's help, can be more obvious to obtain is due to insufficient memory for reasons, but specifically for what reason is not very clear. And most of the similar errors on the Web are clearly cached:
Curl-xput ' Http://localhost:9200/_cache/clear '
After the appeal statement I executed on the server, I clicked on Discover on the web interface of the Kibana and there was an appeal error stating that the idea was wrong, and then I found some useful information in one of the following articles:
http://blog.csdn.net/brotherdong90/article/details/50477836
650) this.width=650; "src=" Https://s4.51cto.com/wyfs02/M02/95/E7/wKioL1ka3qPSB8jEAAD1iwk2iIw579.png "style=" float : none; "title=" 1.png "alt=" Wkiol1ka3qpsb8jeaad1iwk2iiw579.png "/>
650) this.width=650; "src=" Https://s3.51cto.com/wyfs02/M01/95/E7/wKiom1ka3qOBxtf4AACJDrmdWEY064.png "style=" float : none; "title=" 2.png "alt=" Wkiom1ka3qobxtf4aacjdrmdwey064.png "/>
650) this.width=650; "src=" Https://s3.51cto.com/wyfs02/M00/95/E7/wKiom1ka3qShzxv9AAC1G-qPCWo610.png "style=" float : none; "title=" 3.png "alt=" Wkiom1ka3qshzxv9aac1g-qpcwo610.png "/>
In the above we can be relatively clear to get elasticsearch circuite breaker circuit breaker principle, but in the above it appears that the error is fieddate, but my error is request, so I guess I may be because of the request Circuite breaker, request memory is not enough.
650) this.width=650; "src=" Https://s2.51cto.com/wyfs02/M02/95/E7/wKioL1ka4C_yDAwTAACMwVxowQY227.png "title=" 4.png "alt=" Wkiol1ka4c_ydawtaacmwvxowqy227.png "/> The above explanation is primarily to prevent a single request from using too much memory, so what I do is to dynamically change the cache size of the request:
Curl-xput localhost:9200/_cluster/settings-d ' {
"Persistent": {
"Indices.breaker.request.limit": "40%"
}
}‘
The final question was settled satisfactorily.
Elk's Kibana Web error [request] data too large, data for [<agg [2]>] would is larger than limit of