Find what percentage of memcache keys don't have expiries set

find_memcached_expiries.py

import re
import telnetlib
import time

from collections import Counter

hosts = [
    ('localhost', 11211)
]

def get_keys(host, port, sample_ratio=.01, re_key_match=None):
    t = telnetlib.Telnet(host, port)
    t.write('stats items\n')
    items = t.read_until('END').split('\r\n')
    slabs = set([])
    slab_info = {}
    expiries = Counter()
    item_count = 0
    filtered_count = 0
    lowest_timestamp = int(time.mktime(time.gmtime())) + 10000
    for item in items:
        parts = item.split(':')
        if len(parts) == 3:
            slab_id = int(parts[1])
            slabs.add(slab_id)
            if parts[2].split(' ')[0] == 'number':
                slab_info[slab_id] = {}
                slab_info[slab_id]['items'] = int(parts[2].split(' ')[1])

    for slab in slabs:
        sample_count = int(round(slab_info[slab]['items'] * sample_ratio))
        t.write('stats cachedump {0} {1}\n'.format(slab, sample_count))
        cache_lines = t.read_until('END').split('\r\n')

        for line in cache_lines:
            cache_parts = line.split(' ')
            if cache_parts[0] == 'ITEM' and len(cache_parts) == 6:
                if int(cache_parts[4]) < lowest_timestamp:
                    lowest_timestamp = int(cache_parts[4])
                if re_key_match:
                    if re_key_match.match(cache_parts[1]):
                        item_count += 1
                        expiries[cache_parts[4]] += 1
                    else:
                        filtered_count += 1
                else:
                    item_count += 1
                    # Counter for expiries
                    expiries[cache_parts[4]] += 1

    print 'lowest_timestamp: %s' % lowest_timestamp
    no_expiry_count = expiries[str(lowest_timestamp)]

    expiry_ratio = (no_expiry_count / float(item_count)) * 100.0
    if filtered_count > 0:
        print 'Filtered {0} keys due to regex fails'.format(filtered_count)
    print 'Sample ratio of {0}% shows {1} ({2}%) of {3} sampled keys have '\
        'infinite expiries set'.format(
        sample_ratio * 100,
        no_expiry_count,
        expiry_ratio,
        item_count
        )

    print 'Most Common:'
    print expiries.most_common(20)
    t.close()

for host in hosts:
    print 'Host: %s' % host[0]
    # Match keys via an compiled re object
    get_keys(
        host[0], host[1],
        re_key_match=re.compile('^mykeys-.+$')
    )

    # No key matching
    # get_keys(
    #     host[0],
    #     host[1]
    # )