icinga2-checks/check_mysql_slave.py

#!/usr/bin/env python
import argparse
import sys
import traceback

import mysql.connector

from checker import nagios
from checker.result import quit_check
from checker.types import try_int

"""
## MySQL User Setup ##

Attention: The DB-user you type in must have CLIENT REPLICATION rights on the DB-server. Example:
    GRANT REPLICATION CLIENT,SLAVE MONITOR on *.* TO 'nagios'@'monitoringhost' IDENTIFIED BY 'secret';

If you use MariaDB 10.5 or newer, the DB user must have REPLICA MONITOR rights:
    GRANT REPLICA MONITOR ON *.* TO 'nagios'@'monitoringhost' IDENTIFIED BY 'secret';


## Usage example ##

Target delay = 300 seconds
Warning deviation = 10%
Critical deviation = 15%

This will set the warning levels to 330 and 270 seconds, if the delay is greater or less than these values it will return WARNING.
The critical levels will be 345 and 255 seconds.
"""


def main(args):
    cursor = cnx = None
    try:
        cnx = mysql.connector.connect(user=args.username, password=args.password, host=args.host)
        cursor = cnx.cursor(dictionary=True)
        cursor.execute("SHOW SLAVE STATUS")

        slave_status = cursor.fetchone()
        if slave_status is None:
            print("UNKNOWN - Could not retrieve slave status")
            sys.exit(nagios.STATE_UNKNOWN)

        slave_io_running = slave_status['Slave_IO_Running']
        if not slave_io_running:
            print("UNKNOWN - Could not retrieve Slave_IO_Running")
            sys.exit(nagios.STATE_UNKNOWN)

        slave_sql_running = slave_status['Slave_SQL_Running']
        if not slave_sql_running:
            print("UNKNOWN - Could not retrieve Slave_SQL_Running")
            sys.exit(nagios.STATE_UNKNOWN)

        last_io_error = slave_status['Last_IO_Error']
        last_sql_error = slave_status['Last_SQL_Error']

        replication_delay = -1
        perfdata = {}
        exit_code = nagios.STATE_OK
        exit_msg = []

        if slave_io_running != 'Yes':
            exit_code = nagios.STATE_CRIT
            exit_msg.append('Slave IO is not running!')
        if last_io_error:
            exit_code = nagios.STATE_CRIT
            exit_msg.append(f'Last IO Error: {last_io_error.strip(".")}.')
        if slave_sql_running != 'Yes':
            exit_code = nagios.STATE_CRIT
            exit_msg.append('Slave SQL is not running!')
        if last_sql_error:
            exit_code = nagios.STATE_CRIT
            exit_msg.append(f'Last SQL Error: {last_sql_error.strip(".")}.')

        if exit_code == nagios.STATE_OK:
            # Only replication delay if everything else is healthy.
            if not slave_status['Seconds_Behind_Master']:
                quit_check('Slave is not configured for delay', nagios.STATE_CRIT)

            replication_delay = try_int(slave_status['Seconds_Behind_Master'])

            if args.target_delay:
                warn_deviation_max = args.target_delay * (1 + (args.warning_deviation / 100))
                warn_deviation_min = args.target_delay * (1 - (args.warning_deviation / 100))
                crit_deviation_max = args.target_delay * (1 + (args.critical_deviation / 100))
                crit_deviation_min = args.target_delay * (1 - (args.critical_deviation / 100))
                if replication_delay <= crit_deviation_min:
                    exit_code = nagios.STATE_CRIT
                    exit_msg.append('Replication is delayed!')
                elif replication_delay >= crit_deviation_max:
                    exit_code = nagios.STATE_CRIT
                    exit_msg.append('Replication is ahead???')
                elif replication_delay <= warn_deviation_min:
                    exit_code = nagios.STATE_WARN
                    exit_msg.append('Replication is delayed!')
                elif replication_delay >= warn_deviation_max:
                    exit_code = nagios.STATE_WARN
                    exit_msg.append('Replication is ahead???')

        if exit_code == nagios.STATE_OK:
            exit_msg.append('Slave is healthy! Slave SQL: running. Slave IO: running.')

        if replication_delay > -1:
            exit_msg.append(f'Slave is {replication_delay} seconds behind master.')
            perfdata = {
                "replication_delay": {
                    "value": replication_delay,
                    "min": 0,
                    "unit": "s",
                },
            }

        quit_check(' '.join(exit_msg), exit_code, perfdata)

    except mysql.connector.Error as e:
        quit_check(f'Could not connect to database: {e}', nagios.STATE_CRIT)
    except Exception as e:
        quit_check(f'Unknown error: {e}\n{traceback.format_exc()}', nagios.STATE_UNKNOWN)
    finally:
        if cursor:
            cursor.close()
        if cnx:
            cnx.close()


if __name__ == "__main__":
    parser = argparse.ArgumentParser(description='Check a MySQL slave.')
    parser.add_argument('--host', required=True, help='The IP of the slave to connect to.')
    parser.add_argument('--username', required=True, help='Username.')
    parser.add_argument('--password', required=True, help='Password.')
    parser.add_argument('--target-delay', default=None, type=int, help='The target delay in seconds.')
    parser.add_argument('--warning-deviation', default=10, type=int, help='If the delay deviates more than this percentage from the target delay, return warning.')
    parser.add_argument('--critical-deviation', default=15, type=int, help='If the delay deviates more than this percentage from the target delay, return critical.')
    args = parser.parse_args()

    try:
        main(args)
    except Exception as e:
        print(f"UNKNOWN - {e}")
        traceback.print_exc()
        sys.exit(nagios.STATE_UNKNOWN)