icinga2-checks/check_scrutiny_disks.py

192 lines
7.8 KiB
Python
Raw Normal View History

#!/usr/bin/env python3
import argparse
import subprocess
import sys
2023-05-28 18:25:18 -06:00
from datetime import datetime, timedelta
from typing import List
import requests
2023-05-28 14:12:07 -06:00
from checker import nagios
2023-05-28 18:25:18 -06:00
def get_disk_wwn_ids(ignore_non_smart: bool = False) -> List[str] or bool:
wwn_ids = []
try:
2023-05-28 18:25:18 -06:00
output = subprocess.check_output(["sudo", "smartctl", "--scan"])
for line in output.decode("utf-8").strip().split("\n"):
parts = line.split()
2023-05-28 18:25:18 -06:00
if len(parts) >= 3:
device = parts[0]
device_type = parts[2].replace('scsi', 'sat,auto')
try:
smart_supported = subprocess.check_output(["sudo", "smartctl", "-i", device, "-d", device_type]).decode("utf-8")
if "SMART support is: Enabled" in smart_supported:
2023-05-28 18:25:18 -06:00
wwn_line = [line for line in smart_supported.split("\n") if "LU WWN Device Id" in line]
wwn_id = '0x' + wwn_line[0].replace('LU WWN Device Id: ', '').replace(' ', '')
wwn_ids.append(wwn_id)
# else:
# # TODO: warn if a drive doesn't support SMART
except subprocess.CalledProcessError as e:
if ignore_non_smart:
continue
else:
print(f"UNKNOWN: subprocess Error - {e}")
sys.exit(nagios.UNKNOWN)
except subprocess.CalledProcessError as e:
2023-05-28 14:12:07 -06:00
print(f"UNKNOWN: subprocess Error - {e}")
sys.exit(nagios.UNKNOWN)
return wwn_ids
def get_smart_health(wwn_id: str, scrutiny_endpoint: str) -> dict:
url = f"{scrutiny_endpoint}/api/device/{wwn_id}/details"
response = requests.get(url)
if response.status_code == 200:
return response.json()
elif response.status_code == 404:
print(f"Disk {wwn_id} not found on Scrutiny")
return {}
else:
print(f"Scrutiny Error {response.status_code} for disk {wwn_id}: {response.text}")
return {}
2023-05-28 14:12:07 -06:00
def main(args):
results = {}
2023-05-28 18:25:18 -06:00
wwn_ids = get_disk_wwn_ids(args.ignore_non_smart)
for wwn_id in wwn_ids:
2023-05-28 14:12:07 -06:00
smart_health = get_smart_health(wwn_id, args.scrutiny_endpoint)
2023-05-28 18:25:18 -06:00
name = f'/dev/{smart_health["data"]["device"]["device_name"]} {wwn_id}' # differentiate disks in RAID arrays
2023-05-28 14:12:07 -06:00
disk_results = {
'wwn_id': wwn_id,
'failed_attributes': [],
}
metadata = smart_health['metadata']
2023-05-28 18:25:18 -06:00
# For testing
# smart_health['data']['device']['UpdatedAt'] = '2023-04-28T23:00:03.071184465Z'
last_updated = datetime.strptime(smart_health['data']['device']['UpdatedAt'][:-4] + 'Z', '%Y-%m-%dT%H:%M:%S.%fZ')
if datetime.utcnow() - timedelta(hours=args.time_delta_limit) > last_updated:
metics_out_of_date = True
else:
metics_out_of_date = False
2023-05-28 18:25:18 -06:00
if smart_health:
2023-05-28 14:12:07 -06:00
for attribute_id, values in smart_health['data']['smart_results'][0]['attrs'].items():
if values['status'] == 0:
continue
# elif values['status'] == 2 and not args.warn_non_critical:
# continue
values['attribute_name'] = metadata[attribute_id]['display_name']
values['metadata'] = metadata[attribute_id]
2023-05-28 18:25:18 -06:00
2023-05-28 14:12:07 -06:00
if 'observed_thresholds' in values['metadata'].keys():
del values['metadata']['observed_thresholds']
disk_results['failed_attributes'].append(values)
2023-05-28 18:25:18 -06:00
results[name] = disk_results
2023-05-28 14:12:07 -06:00
crit_disks = {}
warn_disks = {}
for disk, values in results.items():
for item in values['failed_attributes']:
if item['status'] == 2 and args.warn_non_critical:
if disk not in warn_disks.keys():
warn_disks[disk] = []
warn_disks[disk].append({
'raw_value': item['raw_value'],
'display_name': item['metadata']['display_name']
})
if item['status'] == 4:
if disk not in crit_disks.keys():
crit_disks[disk] = []
crit_disks[disk].append({
'raw_value': item['raw_value'],
'display_name': item['metadata']['display_name']
})
2023-05-28 14:49:34 -06:00
dt = '<dt>' if args.html else ''
dts = '</dt>' if args.html else ''
dd = '<dd>' if args.html else '\t- '
dds = '</dd>' if args.html else ''
2023-05-28 18:25:18 -06:00
out_of_date_str = f'metrics are >{args.time_delta_limit} hrs out of date ' if metics_out_of_date else ''
2023-05-28 14:12:07 -06:00
return_code = nagios.OK
if len(crit_disks):
return_code = nagios.CRITICAL
2023-05-28 14:49:34 -06:00
if len(warn_disks):
2023-05-28 18:25:18 -06:00
x = f' and {len(warn_disks)} {"warnings" if len(results) > 1 else "warning"}'
2023-05-28 14:49:34 -06:00
else:
x = ''
2023-05-28 18:25:18 -06:00
print(f'CRITICAL: {out_of_date_str + "and " if len(out_of_date_str) else ""}{len(crit_disks)} {"errors" if len(crit_disks) > 1 else "error"}{x}')
2023-05-28 14:49:34 -06:00
print('<dl>')
print(f'{dt}Disks with Errors:{dts}')
2023-05-28 14:12:07 -06:00
for disk, warns in crit_disks.items():
2023-05-28 14:49:34 -06:00
if args.html:
2023-05-28 18:25:18 -06:00
disk_name = f'- <a href="{args.pretty_url}/web/device/{results[disk]["wwn_id"]}" target="_blank">{disk}</a>'
2023-05-28 14:49:34 -06:00
else:
2023-05-28 18:25:18 -06:00
disk_name = f'\t- {disk}'
2023-05-28 14:49:34 -06:00
print(f'{dd}{disk_name}: {", ".join([x["display_name"] for x in warns])}{dds}')
print('</dl>', end='')
2023-05-28 18:25:18 -06:00
if len(out_of_date_str) and not len(crit_disks):
return_code = nagios.CRITICAL
w = "warnings" if len(warn_disks) > 1 else "warning"
print(f'CRITICAL: {out_of_date_str}{"and " + str(len(warn_disks)) + " " + w if len(warn_disks) else ""}')
2023-05-28 14:12:07 -06:00
if len(warn_disks):
if return_code < nagios.CRITICAL:
return_code = nagios.WARNING
2023-05-28 18:25:18 -06:00
print(f'WARNING: {len(warn_disks)} {"warnings" if len(warn_disks) > 1 else "warning"}')
2023-05-28 14:49:34 -06:00
print('<dl>')
print(f'{dt}Disks with issues:{dts}')
2023-05-28 14:12:07 -06:00
for disk, warns in warn_disks.items():
2023-05-28 14:49:34 -06:00
if args.html:
2023-05-28 18:25:18 -06:00
disk_name = f'- <a href="{args.pretty_url}/web/device/{results[disk]["wwn_id"]}" target="_blank">{disk}</a>'
2023-05-28 14:49:34 -06:00
else:
2023-05-28 18:25:18 -06:00
disk_name = f'\t- {disk}'
2023-05-28 14:49:34 -06:00
print(f'{dd}{disk_name}: {", ".join([x["display_name"] for x in warns])}{dds}')
print('</dl>', end='')
2023-05-28 14:12:07 -06:00
if not len(crit_disks) and not len(warn_disks):
2023-05-28 18:25:18 -06:00
print(f'OK: all {len(results.keys())} {"disks" if len(results.keys()) > 1 else "disk"} are healthy!', end='')
2023-05-28 14:12:07 -06:00
2023-05-28 18:25:18 -06:00
print(f"|'warnings'={len(warn_disks)};;; 'errors'={len(crit_disks)};;; 'num_disks'={len(results.keys())};;;")
2023-05-28 14:12:07 -06:00
sys.exit(return_code)
if __name__ == "__main__":
parser = argparse.ArgumentParser(description='')
parser.add_argument('--scrutiny-endpoint', required=True, help='Base URL for scrutiny.')
2023-05-28 18:25:18 -06:00
parser.add_argument('--time-delta-limit', default=24, type=int, help='The Scrutiny data must not be older than this many hours. Default: 24.')
2023-05-28 14:12:07 -06:00
parser.add_argument('--warn-non-critical', action='store_true', help='Warn when a non-critical metric is marked as failed.')
2023-05-28 14:49:34 -06:00
parser.add_argument('--html', action='store_true', help='Print HTML.')
parser.add_argument('--pretty-url', help='The pretty URL to link to when printing HTML.')
2023-05-28 18:25:18 -06:00
parser.add_argument('--ignore-non-smart', action='store_true', help="Ignore any non-SMART devices and any devices that error when reading SMART.")
args = parser.parse_args()
2023-05-28 14:49:34 -06:00
if args.html and not args.pretty_url:
print('UKNOWN: when using --html you must also set --pretty-url')
sys.exit(nagios.UNKNOWN)
args.scrutiny_endpoint = args.scrutiny_endpoint.strip('/')
2023-05-28 14:49:34 -06:00
args.pretty_url = args.pretty_url.strip('/') if args.pretty_url else None
try:
2023-05-28 14:12:07 -06:00
main(args)
except Exception as e:
print(f'UNKNOWN: exception "{e}"')
import traceback
2023-05-28 14:12:07 -06:00
print(traceback.format_exc())
sys.exit(nagios.UNKNOWN)