priority = 60000
retries = 60
-ORDER = ['response_codes', 'response_time', 'requests_per_url', 'http_method', 'bandwidth', 'clients', 'clients_all']
+ORDER = ['response_codes', 'response_time', 'requests_per_url', 'http_method', 'requests_per_ipproto', 'bandwidth', 'clients', 'clients_all']
CHARTS = {
'response_codes': {
'options': [None, 'Response Codes', 'requests/s', 'responses', 'web_log.response_codes', 'stacked'],
['resp_time_avg', 'avg', 'absolute', 1, 1]
]},
'clients': {
- 'options': [None, 'Current Poll Unique Client IPs', 'unique ips', 'unique clients', 'web_log.clients', 'line'],
+ 'options': [None, 'Current Poll Unique Client IPs', 'unique ips', 'unique clients', 'web_log.clients', 'stacked'],
'lines': [
['unique_cur_ipv4', 'ipv4', 'absolute', 1, 1],
['unique_cur_ipv6', 'ipv6', 'absolute', 1, 1]
]},
'clients_all': {
- 'options': [None, 'All Time Unique Client IPs', 'unique ips', 'unique clients', 'web_log.clients_all', 'line'],
+ 'options': [None, 'All Time Unique Client IPs', 'unique ips', 'unique clients', 'web_log.clients_all', 'stacked'],
'lines': [
['unique_tot_ipv4', 'ipv4', 'absolute', 1, 1],
['unique_tot_ipv6', 'ipv6', 'absolute', 1, 1]
'http_method': {
'options': [None, 'Requests Per HTTP Method', 'requests/s', 'requests', 'web_log.http_method', 'stacked'],
'lines': [
+ ]},
+ 'requests_per_ipproto': {
+ 'options': [None, 'Requests Per IP Protocol', 'requests/s', 'requests', 'web_log.requests_per_ipproto', 'stacked'],
+ 'lines': [
+ ['req_ipv4', 'ipv4', 'absolute', 1, 1],
+ ['req_ipv6', 'ipv6', 'absolute', 1, 1]
]}
}
class Service(LogService):
def __init__(self, configuration=None, name=None):
LogService.__init__(self, configuration=configuration, name=name)
- # Vars from module configuration file
+ # Variables from module configuration file
self.log_path = self.configuration.get('path')
self.detailed_response_codes = self.configuration.get('detailed_response_codes', True)
self.all_time = self.configuration.get('all_time', True)
self.url_pattern = self.configuration.get('categories') # dict
- self.regex = None
+ self.regex = None # will be assigned in 'find_regex' method
+ self.resp_time_func = None # will be assigned in 'find_regex' method
+ self._get_data = None # will be assigned in 'check' method.
+ self.order = None # will be assigned in 'create_*_method' method.
+ self.definitions = None # will be assigned in 'create_*_method' method.
+ self.detailed_chart = None # will be assigned in 'create_*_method' method.
+ self.http_method_chart = None # will be assigned in 'create_*_method' method.
# sorted list of unique IPs
self.unique_all_time = list()
# dict for values that should not be zeroed every poll
self.data = {'bytes_sent': 0, 'resp_length': 0, 'resp_time_min': 0,
'resp_time_max': 0, 'resp_time_avg': 0, 'unique_cur_ipv4': 0,
'unique_cur_ipv6': 0, '2xx': 0, '5xx': 0, '3xx': 0, '4xx': 0,
- '1xx': 0, '0xx': 0, 'unmatched': 0}
+ '1xx': 0, '0xx': 0, 'unmatched': 0, 'req_ipv4': 0, 'req_ipv6': 0}
def check(self):
if not self.log_path:
last_line = logs.readline().decode(encoding='utf-8')
# Parse last line
- parsed_line, regex_name = self.find_regex(last_line)
- if not parsed_line:
- self.error('Can\'t parse output')
+ regex_name = self.find_regex(last_line)
+ if not regex_name:
+ self.error('Can\'t parse %s' % self.log_path)
return False
- self.create_charts(parsed_line[0], regex_name)
- if len(parsed_line[0]) == 5:
- self.info('Not all data collected. You need to modify LogFormat.')
- return True
+ if regex_name.startswith('access_'):
+ self.create_access_charts(regex_name)
+ if regex_name == 'access_default':
+ self.info('Not all data collected. You need to modify LogFormat.')
+ self._get_data = self._get_access_data
+ self.info('Used regex: %s' % regex_name)
+ return True
+ else:
+ # If it's not access_logs.. Not used at the moment
+ return False
def find_regex(self, last_line):
+ """
+ :param last_line: str: literally last line from log file
+ :return: regex_name
+ It's sad but different web servers has different logs formats
+ We need to find appropriate regex for current log file
+ All logic is do a regex search through the string for all patterns
+ until we find something or fail.
+ """
# REGEX: 1.IPv4 address 2.HTTP method 3. URL 4. Response code
# 5. Bytes sent 6. Response length 7. Response process time
- default = re.compile(r'([\da-f.:]+)'
- r' -.*?"([A-Z]+)'
- r' (.*?)"'
- r' ([1-9]\d{2})'
- r' (\d+)')
-
- apache_extended = re.compile(r'([\da-f.:]+)'
+ access_default = re.compile(r'([\da-f.:]+)'
r' -.*?"([A-Z]+)'
r' (.*?)"'
r' ([1-9]\d{2})'
- r' (\d+)'
- r' (\d+)'
- r' (\d+) ')
-
- nginx_extended = re.compile(r'([\da-f.:]+)'
- r' -.*?"([A-Z]+)'
- r' (.*?)"'
- r' ([1-9]\d{2})'
- r' (\d+)'
- r' (\d+)'
- r' ([\d.]+) ')
-
- regex_function = zip([apache_extended, nginx_extended, default],
+ r' (\d+)')
+
+ access_apache_ext = re.compile(r'([\da-f.:]+)'
+ r' -.*?"([A-Z]+)'
+ r' (.*?)"'
+ r' ([1-9]\d{2})'
+ r' (\d+)'
+ r' (\d+)'
+ r' (\d+) ')
+
+ access_nginx_ext = re.compile(r'([\da-f.:]+)'
+ r' -.*?"([A-Z]+)'
+ r' (.*?)"'
+ r' ([1-9]\d{2})'
+ r' (\d+)'
+ r' (\d+)'
+ r' ([\d.]+) ')
+
+ regex_function = zip([access_apache_ext, access_nginx_ext, access_default],
[lambda x: x, lambda x: x * 1000, lambda x: x],
- ['apache_extended', 'nginx_extended', 'default'])
-
+ ['access_apache_ext', 'access_nginx_ext', 'access_default'])
+ regex_name = None
for regex, function, name in regex_function:
if regex.search(last_line):
self.regex = regex
self.resp_time_func = function
regex_name = name
break
+ return regex_name
- if self.regex:
- return self.regex.findall(last_line), regex_name
- else:
- return None, None
-
- def create_charts(self, parsed_line, regex_name):
+ def create_access_charts(self, regex_name):
+ """
+ :param regex_name: str: regex name from 'find_regex' method. Ex.: 'apache_extended', 'nginx_extended'
+ :return:
+ Create additional charts depending on the 'find_regex' result (parsed_line) and configuration file
+ 1. 'time_response' chart is removed if there is no 'time_response' in logs.
+ 2. We need to change divisor for 'response_time' chart for apache (time in microseconds in logs)
+ 3. Other stuff is just remove/add chart depending on yes/no in conf
+ """
def find_job_name(override_name, name):
+ """
+ :param override_name: str: 'name' var from configuration file
+ :param name: str: 'job_name' from configuration file
+ :return: str: new job name
+ We need this for dynamic charts. Actually same logic as in python.d.plugin.
+ """
add_to_name = override_name or name
if add_to_name:
- return '_'.join(['web_log', add_to_name])
+ return '_'.join(['web_log', re.sub('\s+', '_', add_to_name)])
else:
return 'web_log'
+ self.order = ORDER[:]
+ self.definitions = deepcopy(CHARTS)
+
job_name = find_job_name(self.override_name, self.name)
self.detailed_chart = 'CHART %s.detailed_response_codes ""' \
' "Response Codes" requests/s responses' \
self.http_method_chart = 'CHART %s.http_method' \
' "" "HTTP Methods" requests/s requests' \
' web_log.http_method stacked 2 %s\n' % (job_name, self.update_every)
- self.order = ORDER[:]
- self.definitions = deepcopy(CHARTS)
- if 'apache' in regex_name:
+
+ if regex_name == 'access_apache_ext':
self.definitions['response_time']['lines'][0][4] = 1000
self.definitions['response_time']['lines'][1][4] = 1000
self.definitions['response_time']['lines'][2][4] = 1000
# Remove 'request_time' chart from ORDER if request_time not in logs
- if len(parsed_line) < 7:
+ if regex_name == 'access_default':
self.order.remove('response_time')
# Remove 'clients_all' chart from ORDER if specified in the configuration
if not self.all_time:
self.order.remove('requests_per_url')
def add_new_dimension(self, dimension, line_list, chart_string, key):
- self.storage.update({dimension: 0})
+ """
+ :param dimension: str: response status code. Ex.: '202', '499'
+ :param line_list: list: Ex.: ['202', '202', 'Absolute']
+ :param chart_string: Current string we need to pass to netdata to rebuild the chart
+ :param key: str: CHARTS dict key (chart name). Ex.: 'response_time'
+ :return: str: new chart string = previous + new dimensions
+ """
+ self.data.update({dimension: 0})
# SET method check if dim in _dimensions
self._dimensions.append(dimension)
# UPDATE method do SET only if dim in definitions
print(chart)
return chart
- def _get_data(self):
+ def _get_access_data(self):
"""
Parse new log lines
- :return: dict
+ :return: dict OR None
+ None if _get_raw_data method fails.
+ In all other cases - dict.
"""
raw = self._get_raw_data()
if raw is None:
default_dict = defaultdict(lambda: 0)
for line in raw:
- match = self.regex.findall(line)
+ match = self.regex.search(line)
if match:
- match_dict = dict(zip_longest('address method url code sent resp_length resp_time'.split(), match[0]))
+ match_dict = dict(zip_longest('address method url code sent resp_length resp_time'.split(),
+ match.groups()))
try:
code = ''.join([match_dict['code'][0], 'xx'])
to_netdata[code] += 1
self._get_data_per_url(match_dict['url'], default_dict)
# requests per http method
self._get_data_http_method(match_dict['method'], default_dict)
-
+ # bandwidth sent
to_netdata['bytes_sent'] += int(match_dict['sent'])
-
+ # request processing time and bandwidth received
if match_dict['resp_length'] and match_dict['resp_time']:
to_netdata['resp_length'] += int(match_dict['resp_length'])
resp_time = self.resp_time_func(float(match_dict['resp_time']))
bisect.insort_left(request_time, resp_time)
request_counter['count'] += 1
request_counter['sum'] += resp_time
+ # requests per ip proto
+ proto = 'ipv4' if '.' in match_dict['address'] else 'ipv6'
+ to_netdata['req_' + proto] += 1
# unique clients ips
if address_not_in_pool(self.unique_all_time, match_dict['address'],
self.storage['unique_tot_ipv4'] + self.storage['unique_tot_ipv6']):
- if '.' in match_dict['address']:
- self.storage['unique_tot_ipv4'] += 1
- else:
- self.storage['unique_tot_ipv6'] += 1
+ self.storage['unique_tot_' + proto] += 1
if address_not_in_pool(unique_current, match_dict['address'],
to_netdata['unique_cur_ipv4'] + to_netdata['unique_cur_ipv6']):
- if '.' in match_dict['address']:
- to_netdata['unique_cur_ipv4'] += 1
- else:
- to_netdata['unique_cur_ipv6'] += 1
+ to_netdata['unique_cur_' + proto] += 1
else:
to_netdata['unmatched'] += 1
# timings
to_netdata['resp_time_min'] = request_time[0]
to_netdata['resp_time_avg'] = float(request_counter['sum']) / request_counter['count']
to_netdata['resp_time_max'] = request_time[-1]
+
to_netdata.update(self.storage)
to_netdata.update(default_dict)
return to_netdata
def _get_data_detailed_response_codes(self, code, default_dict):
- if code not in self.storage:
+ """
+ :param code: str: CODE from parsed line. Ex.: '202, '499'
+ :param default_dict: defaultdict
+ :return:
+ Calls add_new_dimension method If the value is found for the first time
+ """
+ if code not in self.data:
chart_string_copy = self.detailed_chart
self.detailed_chart = self.add_new_dimension(code, [code, code, 'absolute'],
chart_string_copy, 'detailed_response_codes')
default_dict[code] += 1
def _get_data_http_method(self, method, default_dict):
- if method not in self.storage:
+ """
+ :param method: str: METHOD from parsed line. Ex.: 'GET', 'POST'
+ :param default_dict: defaultdict
+ :return:
+ Calls add_new_dimension method If the value is found for the first time
+ """
+ if method not in self.data:
chart_string_copy = self.http_method_chart
self.http_method_chart = self.add_new_dimension(method, [method, method, 'absolute'],
chart_string_copy, 'http_method')
default_dict[method] += 1
def _get_data_per_url(self, url, default_dict):
+ """
+ :param url: str: URL from parsed line
+ :param default_dict: defaultdict
+ :return:
+ Scan through string looking for the first location where patterns produce a match for all user
+ defined patterns
+ """
match = None
for elem in self.url_pattern:
if elem.pattern.search(url):
def address_not_in_pool(pool, address, pool_size):
+ """
+ :param pool: list of ip addresses
+ :param address: ip address
+ :param pool_size: current size of pool
+ :return: True if address not pool and False address in pool
+ If address not in pool function add address to pool.
+ """
index = bisect.bisect_left(pool, address)
if index < pool_size:
if pool[index] == address: