cleanup_zabbix_maildir.py

#!/usr/bin/env python
# encoding: utf-8

'''
Remove Zabbix problem mails if resolved.

This script 
  - recursively searches a maildir
  - looks for mails with PROBLEM in the subject
  - Checks if the proper OK subject exists
  - if so, deletes both mails

Sven Hergenhahn https://github.com/svenXY

'''

import mailbox
import re
import sys
import email.utils
import time
import logging
import operator

class ZabbixMessage(object):
    def __init__(self, subject, date, msgid):
        self.subject = subject
        self.msgid = msgid
        self.date = self.parsetime(date)
        pr = re.compile('PROBLEM:')
        ok = re.compile('OK:')
        if pr.search(self.subject):
            self.typ = 'problem'
        elif ok.search(subject):
            self.typ = 'ok'
        else:
            logging.info('Strange mail: %s' % self.subject)

    def parsetime(self, datestring):
        d = email.utils.parsedate(datestring)
        return int(time.mktime(d))

    def __unicode__(self):
        return 'ZabbixMessage: (%d) %s' % (self.date, self.subject)


def sort_messages_by_date(messages):
    cmpfun = operator.attrgetter("date")
    return messages.sort(key=cmpfun)

def find_problem_msg(messages, subject, date):
    for msg in messages:
        k = re.sub('PROBLEM', 'OK', msg.subject)
        if msg.date < date and subject == k:
            logging.debug('Problem message found for (%d) %s' % (date, subject))
            return msg
    logging.info('No problem message found for (%d) %s' % (date, subject))


if __name__ == '__main__':
    logging.basicConfig(
        level=logging.WARN,
        #filename='/tmp/clean_zabbix_mails.log',
        format='%(asctime)s %(message)s',
        datefmt='%Y%m%d %H:%M:%S')
    try:
        mbox = mailbox.Maildir(sys.argv[1])
    except IndexError:
        logging.fatal('No mailbox specified. Exit.')
        print('Usage: %s /path/to/maildir/containing/zabbixmails' % sys.argv[0])
        sys.exit(1)


    mbox.lock()

    messages = []

    for key,msg in mbox.iteritems():
        messages.append(ZabbixMessage(msg['subject'], msg['date'], key))

    sort_messages_by_date(messages)

    for msg in messages:
        if msg.typ == 'ok':
            ids = [ msg.msgid ]
            prob = find_problem_msg(messages, msg.subject, msg.date)
            try:
                if prob:
                    mbox.discard(prob.msgid)
                    ids.append(prob.msgid)

                mbox.discard(msg.msgid)
                messages = [ m for m in messages if m.msgid not in ids]
            except KeyError as e:
                print('Message not found, thus not deleted: %s' % e)
                logging.error('Message not found, thus not deleted: %s' % e)
            else:
                logging.info('Removed messages: %s' % ', '.join(ids))


    mbox.close()

The latest version sorts the list of messages by date first, compares the timestamps (verifies that the PROBLEM-mail is older) and will only ever delete on PROBLEM-mail per OK-mail.

Some logging has been added as well