Hi,
I use Filebeat with the IIS module to send logs to a distant Logstash, with the following configuration :
Filebeat.yml
filebeat.config.modules:
path: ${path.config}/modules.d/*.yml
# filebeat.inputs:
# other inputs
output.logstash:
hosts: ["ltelsni3.cm-cic.fr:5045"]
compression_level: 3
iis.yml
---
- module: iis
# Access logs
access:
enabled: true
var.paths: ["D:/appli/logfiles/IIS/**/*.log"]
input:
ignore_older: 168h
clean_removed: true
# Error logs
error:
enabled: true
input:
ignore_older: 168h
clean_removed: true
My IIS log files are not archived, resulting in more than 20.000 files in the directory Filebeat is scanning. They all have a similar name (only the date in the name is changing), so I don't have any way to target the most recent ones with the Glob patterns.
When started, Filebeat uses 20% CPU (which is a lot for us), with peaks at 35% CPU when the periodic modifiation date scan occurs.
I know the problem seems obvious : too much files at once. But is there anythin I could have missed (a parameter,...) that could help decreasing the CPU usage ?
Is there a parameter to prevent Filebeat from scanning the modification date of really old files ?
Thanks in advance