The script collects, analyzes and sends a mail notification about filesystem space situation.

space_stat.py

#!/usr/bin/python3
# space_stat.py - collects,  analyses and sends
# a mail notification about filesystem space situation
#
# Author: Andrey Klychkov aaklychkov
# Version: 1.0
# Date: 2018-07-26
# Licence: Copyleft free software
# Requirements: Python3+
#
# Use it by a cron daily job: ./space_stat.py
# If something wrong happens, just clear or fix
# a corresponding stat file.
#
# Configuration:
# 1. Define desired file systems in the FILE_SYSTEMS dict
# 2. Set up your mail server account parameters in
#    the 'Mail params' section
#
# Notification output example:
# ----------------------------
# 2018.04.05 10:34 test-server.lan: Disk space usage statistics
#
# **/mnt/ssd_db**, space usage stat:
# total size: 2.33 TB
# available: 1.18 TB
# one day growth: 93.57 GB
# ------------
# ESTIMATE:
# ------------
# free space will end in 12 day(s) -> 2018.04.17
# ------------
# 7 days average growth: 93.13 GB
# 7 days total growth: 651.93 GB
# 30 days average growth: 93.13 GB
# 30 days total growth: 838.19 GB

import datetime
import math
import os
import smtplib
import socket
import subprocess
import sys
from email.mime.multipart import MIMEMultipart
from email.mime.text import MIMEText


########################
#   PARAMETERS BLOCK   #
########################

# Main params:
# {'/path/to/filesystem': '/path/to/stat/file'}
FILE_SYSTEMS = {'/mnt/ssd_db': '/var/lib/pgsql/andreyk/stat_space_ssd.db',
                '/mnt/hdd_db': '/var/lib/pgsql/andreyk/stat_space_hdd.db'}

# Common params:
HOSTNAME = socket.gethostname()
NOW = datetime.datetime.now()
DATE = NOW.strftime('%Y%m%d')
TIME = NOW.strftime('%Y.%m.%d %H:%M')

# Mail params:
SEND_MAIL = 1
MAIL_SBJ = '%s: Disk space usage statistics' % HOSTNAME
SENDER = 'report.account@example.com'
RECIPIENT = ['mymail@mydomain.com']
SMTP_SRV = 'smtp.gmail.com'
SMTP_PORT = 587
SMTP_PASS = 'pass_of_report.account@_here'


#####################
#   FUCTION BLOCK   #
#####################

def send_mail(sbj, ms):
    if SEND_MAIL:
        msg = MIMEMultipart()
        msg['Subject'] = (sbj)
        msg['From'] = 'root@%s' % HOSTNAME
        msg['To'] = RECIPIENT[0]
        body = MIMEText(ms, 'plain')
        msg.attach(body)
        smtpconnect = smtplib.SMTP(SMTP_SRV, SMTP_PORT)
        smtpconnect.starttls()
        smtpconnect.login(SENDER, SMTP_PASS)
        smtpconnect.sendmail(SENDER, RECIPIENT, msg.as_string())
        smtpconnect.quit()
    else:
        pass


def get_pretty_size(size_bytes):
    if size_bytes == 0:
        return "0B"

    size_name = ("B", "KB", "MB", "GB", "TB", "PB", "EB", "ZB", "YB")
    i = int(math.floor(math.log(size_bytes, 1024)))
    p = math.pow(1024, i)
    s = round(size_bytes / p, 2)
    return "%s %s" % (s, size_name[i])


def db_write(db_file, entry):
    db = open(db_file, 'a')
    try:
        db.write(entry)
    finally:
        db.close()


def db_read(db_file):
    # If there is an abnormal number of rows
    # in a stat file or it has an abnormal size,
    # exit with an error code
    # to prevent out of memory:
    ROWS_LIMIT = 100000

    # Max stat file size in bytes:
    FILE_MAX_SIZE = 104857600

    db_file_size = os.path.getsize(db_file)
    if db_file_size > FILE_MAX_SIZE:
        msg = "Error: size of %s is abnormal: %s" % (
            db_file, get_pretty_size(db_file_size))
        print(msg)
        send_mail(MAIL_SBJ, msg)
        sys.exit(1)

    db = open(db_file, 'r')
    r_count = 0
    for r in db:
        r_count += 1

    if r_count > ROWS_LIMIT:
        msg = "Error: file %s contets too many rows: %s" % (
            db_file, r_count)
        print(msg)
        send_mail(MAIL_SBJ, msg)
        sys.exit(1)

    db.seek(0)

    # Make an array of rows in memory:
    try:
        stat = db.readlines()
        return stat
    finally:
        db.close()


if __name__ == '__main__':
    report = ['%s %s\n\n' % (TIME, MAIL_SBJ)]

    for fs in FILE_SYSTEMS:
        # Get stat about filesystem size:
        statfs = os.statvfs(fs)
        total_fs_size = statfs.f_frsize * statfs.f_blocks
        free_fs_size = statfs.f_frsize * statfs.f_bfree

        db_file = FILE_SYSTEMS.get(fs)
        f_content = db_read(db_file)

        # If a stat file is empty, nothing to compare,
        # write the first string and exit:
        if not f_content:
            entry = '%s,%s,%s\n' % (DATE, free_fs_size, 0)
            db_write(db_file, entry)
            continue

        prev_free_fs_size = int(f_content[-1].split(',')[1])

        if free_fs_size > prev_free_fs_size:
            growth = -(free_fs_size - prev_free_fs_size)
        else:
            growth = (free_fs_size - prev_free_fs_size) * -1

        entry = '%s,%s,%s\n' % (DATE, free_fs_size, growth)

        # Write current stat to a stat file:
        db_write(db_file, entry)

        # Get stat from a file::
        stat = db_read(db_file)

        # Count total amount of growth for last 7 days:
        amount = 0
        for line in stat[-7:]:
            day_growth = int(line.split(',')[2])
            if day_growth > 0:
                amount += day_growth

        # Calculate average 7 days growth:
        if len(stat) < 7:
            average_7growth = amount // len(stat)
        else:
            average_7growth = amount // 7

        # Calculate estimated end space time:
        if average_7growth:
            estim_ttl = int(free_fs_size / average_7growth)
            print(free_fs_size)
            print(average_7growth)
            print(estim_ttl)
            if estim_ttl > 0 and estim_ttl < 3650:
                end_date = (NOW + datetime.timedelta(
                    days=estim_ttl)).strftime('%Y.%m.%d')
            else:
                estim_ttl = '--'
                end_date = 'never'
        else:
            estim_ttl = '--'
            end_date = 'never'

        tot_growth7 = get_pretty_size(amount)
        av_growth7 = get_pretty_size(average_7growth)

        # Calculate average 30 days growth:
        if len(stat) > 7:
            amount = 0
            for line in stat[-30:]:
                day_growth = int(line.split(',')[2])
                if day_growth > 0:
                    amount += day_growth

            if len(stat) < 30:
                average_30growth = amount / len(stat)
            else:
                average_30growth = amount / 30

            tot_growth30 = get_pretty_size(amount)
            av_growth30 = get_pretty_size(average_30growth)
        else:
            tot_growth30 = 'no data'
            av_growth30 = 'no data'

        # Make a part of the mail report for a specific fs:
        total_size = get_pretty_size(total_fs_size)
        available = get_pretty_size(free_fs_size)

        if growth > 0:
            pretty_growth = get_pretty_size(growth)
        else:
            pretty_growth = '0 B'

        sep = ('-' * 12) + '\n'
        msg = ['**%s**, space usage stat:\n' % fs,
               'total size: %s\n' % total_size,
               'available: %s\n' % available,
               'one day growth: %s\n' % pretty_growth,
               sep,
               'ESTIMATE:\n',
               sep,
               'free space will end in %s day(s) -> %s\n' % (
                   estim_ttl, end_date),
               sep,
               '7 days average growth: %s\n' % av_growth7,
               '7 days total growth: %s\n' % tot_growth7,
               '30 days average growth: %s\n' % av_growth30,
               '30 days total growth: %s\n\n' % tot_growth30]

        for i in msg:
            report.append(i)

    for i in report:
        print(i, end='')

    # Make a body of the mail report:
    report_msg = ''.join(report)

    # Send the mail report:
    send_mail(MAIL_SBJ, report_msg)

    sys.exit(0)