Ansible Log Scraper

ansible-log-analyzer.py

#!/usr/bin/env python
#
# Usage: ansible-log-analyzer.py [LOGFILE]

import re
import sys
from datetime import datetime, timedelta

TASK_RE = re.compile(r'(\d{4}-\d{2}-\d{2} \d{2}:\d{2}:\d{2},\d{3}) p=\d+ u=[\w-]* \|\s*TASK: \[(.*)\]')

first_task_time = None

last_task_name = None
last_task_start = None

# Map task name to a dict of 'iterations' (an int) and 'runtime' (a timedelta):
results = {}

with open(sys.argv[1]) as f:
    for line in f:
        match = TASK_RE.match(line)
        if match:
            task_name = match.groups()[1]
            task_start = datetime.strptime(match.groups()[0], '%Y-%m-%d %H:%M:%S,%f')

            if task_name not in results:
                results[task_name] = {
                    "iterations": 0,
                    "runtime": timedelta(),
                }

            if first_task_time is None:
                first_task_time = task_start
            else:
                # First adjust the interval and runtime for previous task
                delta = task_start - last_task_start
                results[last_task_name]["iterations"] = results[last_task_name]["iterations"] + 1
                results[last_task_name]["runtime"] = results[last_task_name]["runtime"] + delta
            last_task_name = task_name
            last_task_start = task_start

def left_pad(i):
    if i < 10:
        return "0%s" % i
    return "%s" % i

# Sort the results into a list of tuples, descending order based on total runtime:
tuples = [(k, v['iterations'], v['runtime']) for k, v in results.iteritems()]
sorted_tuples = sorted(tuples, key=lambda x: x[2])
for task, iterations, runtime in sorted_tuples:
    print task
    print "   iterations: %s" % iterations
    #hours, remainder = divmod(runtime.seconds, 3600)
    minutes, seconds = divmod(runtime.seconds, 60)
    print "   total time: %s:%s" % (left_pad(minutes), left_pad(seconds))