Skip to main content

Network and Disk I/O Stats Monitor

Project description

iometrics

Python

Monitor and log Network and Disks statistics in MegaBytes per second.

Install

pip install iometrics

Usage

Pytorch-lightning integration

from pytorch_lightning import Trainer
from iometrics.pytorch_lightning.callbacks import NetworkAndDiskStatsMonitor

net_disk_stats = NetworkAndDiskStatsMonitor()

trainer = Trainer(callbacks=[net_disk_stats])

Metrics generated

  • network/recv_MB_per_sec – Received MB/s on all relevant network interfaces as a SUM.
  • network/sent_MB_per_sec – Sent MB/s on all relevant network interfaces as a SUM.
  • disk/util% – Disk utilization percentage as the average of all disk devices.
  • disk/read_MB_per_sec – Disks read MB/s as the sum of all disk devices.
  • disk/writ_MB_per_sec – Disks written MB/s as the sum of all disk devices.
  • disk/io_read_count_per_sec – Disks read I/O operations per second as the sum of all disk devices.
  • disk/io_writ_count_per_sec – Disks written I/O operations per second as the sum of all disk devices.

Screen shot

Pure-Python implementation (zero dependencies)

Quick check

python -c 'from iometrics.example import usage; usage()'

Example output

|        Network (MBytes/s)       | Disk Util |            Disk MBytes          |           Disk I/O          |
|     Received    |     Sent      |     %     |    MB/s Read    |  MB/s Written |     I/O Read    | I/O Write |
|   val  |   avg  |  val  |  avg  | val | avg |  val   |  avg   |  val  |  avg  |   val  |   avg  | val | avg |
| ------:| ------:| -----:| -----:| ---:| ---:| ------:| ------:| -----:| -----:| ------:| ------:| ---:| ---:|
|    4.6 |    3.5 |   0.1 |   0.1 |  49 |   2 |   52.8 |    1.1 |   0.0 |   0.9 |    211 |      4 |   5 |  18 |
|    4.1 |    3.5 |   0.1 |   0.1 |  61 |   3 |   60.4 |    2.4 |  40.3 |   1.7 |    255 |     10 | 149 |  21 |

Full code

import time
from iometrics import NetworkMetrics, DiskMetrics
from iometrics.example import DUAL_METRICS_HEADER
net  = NetworkMetrics()
disk = DiskMetrics()
for i in range(100):
    time.sleep(1)
    net.update_stats()
    disk.update_stats()
    if i % 15 == 0:
        print(DUAL_METRICS_HEADER)
    row = (
        f"| {net.mb_recv_ps.val:6.1f} | {net.mb_recv_ps.avg:6.1f} "
        f"| {net.mb_sent_ps.val:5.1f} | {net.mb_sent_ps.avg:5.1f} "
        f"| {int(disk.io_util.val):3d} | {int(disk.io_util.avg):3d} "
        f"| {disk.mb_read.val:6.1f} | {disk.mb_read.avg:6.1f} "
        f"| {disk.mb_writ.val:5.1f} | {disk.mb_writ.avg:5.1f} "
        f"| {int(disk.io_read.val):4d} | {int(disk.io_read.avg):4d} "
        f"| {int(disk.io_writ.val):3d} | {int(disk.io_writ.avg):3d} "
        f"|"
    )
    print(row)

Run in a Docker container

Containers don't have access to the host's network statistics, therefore this workaround is needed.

# on the host machine (not inside the container)
iometrics replicate proc &

after you run above script in the host you should mount the /host/proc/net/dev to the container. for example:

docker run -it -v "/tmp/proc_net_dev:/host/proc/net/dev:ro" <YOURIMAGE>

Contributing

See CONTRIBUTING.md

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

iometrics-0.0.8.tar.gz (14.5 kB view hashes)

Uploaded Source

Built Distribution

iometrics-0.0.8-py3-none-any.whl (16.6 kB view hashes)

Uploaded Python 3

Supported by

AWS AWS Cloud computing and Security Sponsor Datadog Datadog Monitoring Fastly Fastly CDN Google Google Download Analytics Microsoft Microsoft PSF Sponsor Pingdom Pingdom Monitoring Sentry Sentry Error logging StatusPage StatusPage Status page