Reputation: 13
i want to ask or maybe we can discuss together.
So i use fluentbit as a log scapper on my kubernetes cluster, but when in backend service have a lot of process, fluentbit will work faster, and the resources used are higher for scrape logs and sometimes nodes often die suddenly.
How to solve this problem? because I want to fluentbit working well when trafic process high
Notes: this is my fluentbit input config
Buffer_Chunk_Size 150k
Buffer_Max_Size 150k
Mem_Buf_Limit 5MB
Skip_Long_Lines Off
Skip_Empty_Lines On
Refresh_Interval 10
Read_from_head true
Thanks for your advice!
Upvotes: 1
Views: 1227
Reputation: 416
I would suggest:
Buffer_Chunk_Size 1M
Buffer_Max_Size 1M
and if you are using v2.0 add:
threaded on
also make sure that your flush interval is set to 1
second (SERVICE section)
Upvotes: 2