Skip to content

Review system integration default metrics collection frequency #6825

Open

Description

Hi,

We saw the system integration metrics causing performance issues on multiple linux production servers.

The default settings are collection data too regularly. Every 60-120 seconds is okay, not every 10 seconds. The integration was using 30% CPU on production servers.

System socket_summary metrics and System network metrics need reducing to every 5 minutes. These were causing most of the CPU. The CPU usage reduces to 15% when we increased the time to 5 minutes.

IT weren't happy.

Thanks

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Metadata

Assignees

No one assigned

    Labels

    Integration:systemSystemTeam:Obs-InfraObsLabel for the Observability Infrastructure Monitoring team [elastic/obs-infraobs-integrations]

    Type

    No type

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions