Greetings
I am seeing a lot of lines in my Elasticsearch logs which look like this:
[2018-11-02T08:00:12,056][INFO ][o.e.m.j.JvmGcMonitorService] [vie01a-clog-pesd-pe05] [gc][24] overhead, spent [277ms] collecting in the last [1s]
[2018-11-02T08:00:25,212][INFO ][o.e.m.j.JvmGcMonitorService] [vie01a-clog-pesd-pe05] [gc][37] overhead, spent [355ms] collecting in the last [1.1s]
[2018-11-02T08:01:12,534][INFO ][o.e.m.j.JvmGcMonitorService] [vie01a-clog-pesd-pe05] [gc][84] overhead, spent [374ms] collecting in the last [1.2s]
[2018-11-02T08:01:42,570][INFO ][o.e.m.j.JvmGcMonitorService] [vie01a-clog-pesd-pe05] [gc][114] overhead, spent [432ms] collecting in the last [1s]
[2018-11-02T08:01:58,697][INFO ][o.e.m.j.JvmGcMonitorService] [vie01a-clog-pesd-pe05] [gc][130] overhead, spent [304ms] collecting in the last [1.1s]
[2018-11-02T08:02:15,709][INFO ][o.e.m.j.JvmGcMonitorService] [vie01a-clog-pesd-pe05] [gc][147] overhead, spent [337ms] collecting in the last [1s]
[2018-11-02T08:02:28,096][INFO ][o.e.m.j.JvmGcMonitorService] [vie01a-clog-pesd-pe05] [gc][159] overhead, spent [431ms] collecting in the last [1.3s]
[2018-11-02T08:03:21,184][INFO ][o.e.m.j.JvmGcMonitorService] [vie01a-clog-pesd-pe05] [gc][212] overhead, spent [426ms] collecting in the last [1s]
[2018-11-02T08:14:45,859][INFO ][o.e.m.j.JvmGcMonitorService] [vie01a-clog-pesd-pe05] [gc][895] overhead, spent [590ms] collecting in the last [1.6s]
[2018-11-02T08:17:35,096][WARN ][o.e.m.j.JvmGcMonitorService] [vie01a-clog-pesd-pe05] [gc][1064] overhead, spent [643ms] collecting in the last [1s]
My cluster is ingesting about 5000 events per second. It has 6 nodes, each has 96 GB of ram, where 30 GB is dedicated for heap (-Xms30g
and -Xmx30g
).
I see no adverse impact on the performance, and I would like to increase the logging threshold from INFO
to WARN
, so that the garbage collector doesn't spam my log files.
I think the command will look something like
PUT /_cluster/settings
{
"persistent": {
"logger.org.elasticsearch.something.something.JvmGcMonitorService": "WARN"
}
}
However, I do not know the exact logger path, since in the logs it is abbreviated to o.e.m.j.JvmGcMonitorService
, and I can't find the relevant documentation.
What logger path should I use? Alternatively, how should I address this issue?