Did you add up all the RAM that was given to Elasticsearch/Graylog.
Looks like you have 9GB RAM on that node.
So if you have 2Gb of heap for Graylog, 4Gb of Heap for elasticsearch that would leave 3 Gb for the system. Is this correct?
EDIT: I just realized, This node only has Graylog/MongoDb.
It might be, not sure. I assume this is new node? and what are you ingest rate for messages?
It took a minute , I see now. You gave Graylog 4 Gb of heap and it look like its using 7 and perhaps 2 left for the system.
Out of curiosity , if you add let’s say 4 more GB of RAM does it still increase past 7 GB mark?
I noticed a pattern that when I turn off the old mongo cluster (which I combined with the new one), lags begin. If there are less than 3 mongo servers, then it starts to slow down.
Hello,
This does make sense, thou. Specially for the amount of logs you ingesting.
Looks like Graylog is using 6 GB out of 10.
Is it still increasing? or is it staying around 6 GB give or take 1 GB?
now it costs 60 GB ± 1
Hello. Nothing new. After making graylog with mongo separately. it seems that the RAM has stopped filling up and is kept around 60 ± 10% of 10 GB. So basically I think we won.
This topic was automatically closed 14 days after the last reply. New replies are no longer allowed.
