-
Notifications
You must be signed in to change notification settings - Fork 5k
Open
Labels
needs_teamIndicates that the issue/PR needs a Team:* labelIndicates that the issue/PR needs a Team:* label
Description
Hey, We Use Filebeat Version 3.17.5
The Filebeat collects logs and sends to Kafka;
The config is
filebeat.inputs:
- type: log
path: [cd /xxxxx]
fields:
ip: "xxx.xxx.xxx.xx"
system: "xxxx"
logType: 'link'
logSubType: -1
fields_under_root: true
include_lines: [ERROR]
exclude_lines:[INFO]
monitorining.enabled: true
output.kafka.enabled: true
output.kafka.hosts: ['xxx.xx.xx.xx:9092']
output.kafka.topic: '%{[system]}-%{[logType]}%{[logSubType]}'
output.kafka.compression: gzip
output.kafka.max_message_types: 100000
output.kafka.required_acks: 1
logging.to_files: true
close_inactive: 5m
close_timeout: 2m
clean_inactive: 12h
close_older: 1m
close_renamed: true
close_removed: true
tail_files: true
max_procs: 1
Filebeat's normally memory usage is around 400MB
However, there was a unexcpected surge in memory consumption one day.
the Filebeat's memery used about:
15:25:55 436M
15:26:26 776M
15:26:58 10.6G
15:27:29 12.7G
15:28:01 15.9G
15:28:33 19.2G
....
15:38:20 55.2G
then the Filebeat oom killed
To begin with, the amount of logs is staying below 1 GB
and there are no abnormal log messages (including multiline log entries) in the logs.
Since no logs were printed and the issue cannot be reproduced at the moment, please assist in analyzing it.
Reactions are currently unavailable
Metadata
Metadata
Assignees
Labels
needs_teamIndicates that the issue/PR needs a Team:* labelIndicates that the issue/PR needs a Team:* label