Reduce chaperone Load by Pruning Special FS Types from find Searches

[arno01]

The chaperone utility employs the command find / -type f ... to search through pods.

This search inadvertently includes non-filesystem-based paths such as /proc, /sys, etc.

Such inclusion not only generates additional processing load but can also lead to issues. I recall instances from a few years ago when the find / command would get indefinitely stuck.

I recommend using the command find / \( -fstype proc -o -fstype sysfs -o -fstype tmpfs \) -prune -o -type f to specifically exclude special filesystem types like proc, sysfs, and tmpfs.

In this instance, a single find command unnecessarily scans an additional 62,201 files.

@88plug

[arno01]

I'm not certain about the frequency of the operation - whether it's every 1 or 5 minutes - but I would suggest extending the interval to at least 10 minutes.

I would generally recommend using inotifywatch, as it can eliminate the need to repeatedly scan all the data.

[88plug]

I recommend using the command find / \( -fstype proc -o -fstype sysfs -o -fstype tmpfs \) -prune -o -type f to specifically exclude special filesystem types like proc, sysfs, and tmpfs.

This was in some initial releases, go ahead and propose the change.

[88plug]

I'm not certain about the frequency of the operation - whether it's every 1 or 5 minutes - but I would suggest extending the interval to at least 10 minutes.

if __name__ == "__main__":
    parser = argparse.ArgumentParser(description='Check for issues in Kubernetes pods.')
    parser.add_argument('--delete-namespaces', action='store_true', help='Delete namespaces that have issues.')
    parser.add_argument('--discord', action='store_true', help='Send notifications to Discord when deleting namespaces.')
    parser.add_argument('--pushover', action='store_true', help='Send notifications to Pushover when deleting namespaces.')
    parser.add_argument('--time', type=str, default="", help='Run the application for a given time duration. Examples: --1h, --10m, --forever.')

    args = parser.parse_args()

    time_mapping = {
        'forever': None,
        '1h': 3600,
        '10m': 600,
        '1m': 60,
        '30s': 30,
        # Add more mappings here
    }

is how it works, you use arguments to control Chaperone.

[88plug]

I would generally recommend using inotifywatch, as it can eliminate the need to repeatedly scan all the data.

The app watches pods, not data.

[andy108369]

I would generally recommend using inotifywatch, as it can eliminate the need to repeatedly scan all the data.

The app watches pods, not data.

"Watches the pods" implies scanning their data ("data" - in the context of the data paths, i.e. the directories, filenames; obviously not the contents).