check that a site works, but do not fail from the first attempt

check_url.py

#!/usr/bin/env python

import argparse
import hashlib
import os
import traceback
import urllib2
import sys

def try_url(url, retries=1):
    while retries:
        retries -= 1
        try:
            urllib2.urlopen(url).read()
        except Exception, e:
            if not retries:
                raise
        else:
            return

USAGE = """
check_url.py checks if a url works, and exist with 1, if it does not. However, the main
feature is checking multiple times so that a single failure does not count.

E.g.

    check_url.py http://example.com/ --instant-retries 2 --deferred-retries 1

will try to get the url 2 times, and if it's not the first time the script is launched,
it will fail with an exception.
"""

if __name__ == '__main__':
    parser = argparse.ArgumentParser()
    parser.add_argument('url')
    parser.add_argument('--instant-retries', type=int, default=2)
    parser.add_argument('--deferred-retries', type=int, default=1)
    parser.add_argument('--storage-dir', default='/tmp/.check_url/')
    args = parser.parse_args()

    if not os.path.exists(args.storage_dir):
        os.makedirs(args.storage_dir)

    attempts_path = os.path.join(args.storage_dir, 'attempts_%s' % hashlib.md5(args.url).hexdigest())

    try:
        with open(attempts_path) as f:
            dattempts = int(f.read().strip())
    except IOError:
        dattempts = 0

    # print args.url, args.instant_retries, args.deferred_retries

    e = None

    try:
        try_url(args.url, args.instant_retries)
    except Exception, e:
        dattempts += 1
        with open(attempts_path, 'w') as f:
            f.write(str(dattempts))
        if dattempts > args.deferred_retries:
            traceback.print_exc()
            sys.exit(1)
    else:
        try:
            os.unlink(attempts_path)
        except OSError:
            pass