retries = 60
ORDER = ['response_statuses', 'response_codes', 'bandwidth', 'response_time', 'requests_per_url', 'http_method',
- 'requests_per_ipproto', 'clients', 'clients_all']
+ 'http_version', 'requests_per_ipproto', 'clients', 'clients_all']
CHARTS = {
'response_codes': {
'options': [None, 'Response Codes', 'requests/s', 'responses', 'web_log.response_codes', 'stacked'],
'http_method': {
'options': [None, 'Requests Per HTTP Method', 'requests/s', 'http methods', 'web_log.http_method', 'stacked'],
'lines': [
+ ['GET', 'GET', 'incremental', 1, 1]
]},
+ 'http_version': {
+ 'options': [None, 'Requests Per HTTP Version', 'requests/s', 'http versions',
+ 'web_log.http_version', 'stacked'],
+ 'lines': []},
'requests_per_ipproto': {
'options': [None, 'Requests Per IP Protocol', 'requests/s', 'ip protocols', 'web_log.requests_per_ipproto',
'stacked'],
# 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.
"""
LogService.__init__(self, configuration=configuration, name=name)
# Variables from module configuration file
+ self.type = self.configuration.get('type', 'web_access')
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.custom_log_format = self.configuration.get('custom_log_format') # dict
# Instance variables
- self.unique_all_time = list() # sorted list of unique IPs
self.regex = None # will be assigned in 'find_regex' or 'find_regex_custom' method
- self.resp_time_func = None # will be assigned in 'find_regex' or 'find_regex_custom' method
- # if there is no new logs this dict returned to netdata
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, 'req_ipv4': 0,
'req_ipv6': 0, 'unique_tot_ipv4': 0, 'unique_tot_ipv6': 0, 'successful_requests': 0,
- 'redirects': 0, 'bad_requests': 0, 'server_errors': 0, 'other_requests': 0}
+ 'redirects': 0, 'bad_requests': 0, 'server_errors': 0, 'other_requests': 0, 'GET': 0}
def check(self):
+ """
+ :return: bool
+
+ 1. "log_path" is specified in the module configuration file
+ 2. "log_path" must be readable by netdata user and must exist
+ 3. "log_path' must not be empty. We need at least 1 line to find appropriate pattern to parse
+ 4. other checks depends on log "type"
+ """
if not self.log_path:
self.error('log path is not specified')
return False
- # log_path must be readable
if not access(self.log_path, R_OK):
self.error('%s not readable or not exist' % self.log_path)
return False
- # log_path file should not be empty
if not getsize(self.log_path):
self.error('%s is empty' % self.log_path)
return False
self.error(str(error))
return False
- # Custom_log_format is preferable
- if self.custom_log_format:
- match_dict, log_name, error = self.find_regex_custom(last_line)
- else:
- match_dict, log_name, error = self.find_regex(last_line)
+ if self.type == 'web_access':
+ self.unique_all_time = list() # sorted list of unique IPs
+ self.detailed_response_codes = self.configuration.get('detailed_response_codes', True)
+ self.all_time = self.configuration.get('all_time', True)
- if match_dict is None:
- self.error(str(error))
- return False
+ # Custom_log_format or predefined log format.
+ if self.custom_log_format:
+ match_dict, error = self.find_regex_custom(last_line)
+ else:
+ match_dict, error = self.find_regex(last_line)
- if not (self.regex and self.resp_time_func):
- self.error('That can not happen, but it happened. "regex" or "resp_time_func" is None')
+ # "match_dict" is None if there are any problems
+ if match_dict is None:
+ self.error(str(error))
+ return False
+
+ # self.url_pattern check
+ if self.url_pattern:
+ self.url_pattern = check_req_per_url_pattern('rpu', self.url_pattern)
- if log_name == 'web_access':
self.create_access_charts(match_dict) # Create charts
- self._get_data = self._get_access_data
- self.info('Collected data: %s' % list(match_dict.keys()))
- return True
+ self._get_data = self._get_access_data # _get_data assignment
else:
- # If it's not access_logs.. Not used at the moment
+ self.error('Not implemented')
return False
+ # Double check
+ if not self.regex:
+ self.error('That can not happen, but it happened. "regex" is None')
+
+ self.info('Collected data: %s' % list(match_dict.keys()))
+ return True
+
def find_regex_custom(self, last_line):
"""
:param last_line: str: literally last line from log file
:return: tuple where:
[0]: dict or None: match_dict or None
- [1]: str or None: log_name or None
- [2]: str: error description
+ [1]: str: error description
We are here only if "custom_log_format" is in logs. We need to make sure:
1. "custom_log_format" is a dict
("resp_length" is integer or "-", "resp_time" is integer or float)
"""
- try:
- self.custom_log_format.keys()
- except AttributeError:
- return None, None, 'Custom log: "custom_log_format" is not a <dict>'
+ if not is_dict(self.custom_log_format):
+ return find_regex_return(msg='Custom log: "custom_log_format" is not a <dict>')
pattern = self.custom_log_format.get('pattern')
if not (pattern and isinstance(pattern, str)):
- return None, None, 'Custom log: "pattern" option is not specified or type is not <str>'
+ return find_regex_return(msg='Custom log: "pattern" option is not specified or type is not <str>')
resp_time_func = self.custom_log_format.get('time_multiplier') or 0
if not isinstance(resp_time_func, int):
- return None, None, 'Custom log: "time_multiplier" is not an integer'
+ return find_regex_return(msg='Custom log: "time_multiplier" is not an integer')
+
+ try:
+ regex = re.compile(pattern)
+ except re.error as error:
+ return find_regex_return(msg='Pattern compile error: %s' % str(error))
- regex = re.compile(pattern)
match = regex.search(last_line)
if match:
match_dict = match.groupdict() or None
else:
- return None, None, 'Custom log: pattern search FAILED'
+ return find_regex_return(msg='Custom log: pattern search FAILED')
if match_dict is None:
- return None, None, 'Custom log: search OK but contains no named subgroups' \
- ' (you need to use ?P<subgroup_name>)'
+ find_regex_return(msg='Custom log: search OK but contains no named subgroups'
+ ' (you need to use ?P<subgroup_name>)')
else:
- basic_values = {'address', 'method', 'url', 'code', 'bytes_sent'} - set(match_dict)
-
- if basic_values:
- return None, None, 'Custom log: search OK but some mandatory keys (%s) are missing' % list(basic_values)
+ mandatory_dict = {'address': r'[\da-f.:]+',
+ 'code': r'[1-9]\d{2}',
+ 'method': r'[A-Z]+',
+ 'bytes_sent': r'\d+|-'}
+ optional_dict = {'resp_length': r'\d+',
+ 'resp_time': r'[\d.]+',
+ 'http_version': r'\d\.\d'}
+
+ mandatory_values = set(mandatory_dict) - set(match_dict)
+ if mandatory_values:
+ return find_regex_return(msg='Custom log: search OK but some mandatory keys (%s) are missing'
+ % list(mandatory_values))
+ else:
+ for key in mandatory_dict:
+ if not re.search(mandatory_dict[key], match_dict[key]):
+ return find_regex_return(msg='Custom log: can\'t parse "%s": %s'
+ % (key, match_dict[key]))
+
+ optional_values = set(optional_dict) & set(match_dict)
+ for key in optional_values:
+ if not re.search(optional_dict[key], match_dict[key]):
+ return find_regex_return(msg='Custom log: can\'t parse "%s": %s'
+ % (key, match_dict[key]))
+
+ dot_in_time = '.' in match_dict.get('resp_time', '')
+ if dot_in_time:
+ self.resp_time_func = lambda time: time * (resp_time_func or 1000000)
else:
- if not re.search(r'[\da-f.:]+', match_dict['address']):
- return None, None, 'Custom log: can\'t parse "address": %s' % match_dict['address']
- if not re.search(r'[1-9]\d{2}', match_dict['code']):
- return None, None, 'Custom log: can\'t parse "code": %s' % match_dict['code']
- if not re.search(r'[A-Z]+', match_dict['method']):
- return None, None, 'Custom log: can\'t parse "method": %s' % match_dict['method']
- if not re.search(r'\d+|-', match_dict['bytes_sent']):
- return None, None, 'Custom log: can\'t parse "bytes_sent": %s' % match_dict['bytes_sent']
-
- if 'resp_length' in match_dict:
- if not re.search(r'\d+', match_dict.get('resp_length', '')):
- return None, None, 'Custom log: can\'t parse "resp_length": %s' % match_dict['resp_length']
-
- if 'resp_time' in match_dict:
- if not re.search(r'[\d.]+', match_dict.get('resp_length', '')):
- return None, None, 'Custom log: can\'t parse "resp_time": %s' % match_dict['resp_time']
- else:
- if '.' in match_dict['resp_time']:
- self.resp_time_func = lambda time: time * (resp_time_func or 1000000)
- else:
- self.resp_time_func = lambda time: time * (resp_time_func or 1)
+ self.resp_time_func = lambda time: time * (resp_time_func or 1)
self.regex = regex
- return match_dict, 'web_access', 'Custom log: we are fine'
+ return find_regex_return(match_dict=match_dict)
def find_regex(self, last_line):
"""
:param last_line: str: literally last line from log file
:return: tuple where:
[0]: dict or None: match_dict or None
- [1]: str or None: log_name or None
- [2]: str: error description
+ [1]: str: error description
We need to find appropriate pattern for current log file
All logic is do a regex search through the string for all predefined patterns
until we find something or fail.
# 5. Bytes sent 6. Response length 7. Response process time
acs_default = re.compile(r'(?P<address>[\da-f.:]+)'
r' -.*?"(?P<method>[A-Z]+)'
- r' (?P<url>.*?)"'
+ r' (?P<url>[^ ]+)'
+ r' [A-Z]+/(?P<http_version>\d\.\d)"'
r' (?P<code>[1-9]\d{2})'
r' (?P<bytes_sent>\d+|-)')
acs_apache_ext_insert = re.compile(r'(?P<address>[\da-f.:]+)'
r' -.*?"(?P<method>[A-Z]+)'
- r' (?P<url>.*?)"'
+ r' (?P<url>[^ ]+)'
+ r' [A-Z]+/(?P<http_version>\d\.\d)"'
r' (?P<code>[1-9]\d{2})'
r' (?P<bytes_sent>\d+|-)'
r' (?P<resp_length>\d+)'
acs_apache_ext_append = re.compile(r'(?P<address>[\da-f.:]+)'
r' -.*?"(?P<method>[A-Z]+)'
- r' (?P<url>.*?)"'
+ r' (?P<url>[^ ]+)'
+ r' [A-Z]+/(?P<http_version>\d\.\d)"'
r' (?P<code>[1-9]\d{2})'
r' (?P<bytes_sent>\d+|-)'
r' .*?'
acs_nginx_ext_insert = re.compile(r'(?P<address>[\da-f.:]+)'
r' -.*?"(?P<method>[A-Z]+)'
- r' (?P<url>.*?)"'
+ r' (?P<url>[^ ]+)'
+ r' [A-Z]+/(?P<http_version>\d\.\d)"'
r' (?P<code>[1-9]\d{2})'
r' (?P<bytes_sent>\d+)'
r' (?P<resp_length>\d+)'
acs_nginx_ext_append = re.compile(r'(?P<address>[\da-f.:]+)'
r' -.*?"(?P<method>[A-Z]+)'
- r' (?P<url>.*?)"'
+ r' (?P<url>[^ ]+)'
+ r' [A-Z]+/(?P<http_version>\d\.\d)"'
r' (?P<code>[1-9]\d{2})'
r' (?P<bytes_sent>\d+)'
r' .*?'
match_dict = match.groupdict()
break
- return match_dict or None, 'web_access', 'Unknown log format. Plugin still can work for you.' \
- ' Read about the "custom_log_format" feature in the conf file'
+ return find_regex_return(match_dict=match_dict or None,
+ msg='Unknown log format. You need to use "custom_log_format" feature.')
def create_access_charts(self, match_dict):
"""
' web_log.detailed_response_codes stacked 1 %s\n' % (job_name, self.update_every)
self.http_method_chart = 'CHART %s.http_method' \
' "" "Requests Per HTTP Method" requests/s "http methods"' \
- ' web_log.http_method stacked 2 %s\n' % (job_name, self.update_every)
+ ' web_log.http_method stacked 2 %s\n' \
+ 'DIMENSION GET GET incremental\n' % (job_name, self.update_every)
+ self.http_version_chart = 'CHART %s.http_version' \
+ ' "" "Requests Per HTTP Version" requests/s "http versions"' \
+ ' web_log.http_version stacked 3 %s\n' % (job_name, self.update_every)
# Remove 'request_time' chart from ORDER if resp_time not in match_dict
if 'resp_time' not in match_dict:
# Add 'requests_per_url' chart if specified in the configuration
if self.url_pattern:
- self.url_pattern = [NAMED_URL_PATTERN(description=k, pattern=re.compile(v)) for k, v
- in self.url_pattern.items()]
self.definitions['requests_per_url'] = {'options': [None, 'Requests Per Url', 'requests/s',
'urls', 'web_log.requests_per_url', 'stacked'],
- 'lines': [['other_url', 'other', 'incremental']]}
+ 'lines': [['rpu_other', 'other', 'incremental']]}
for elem in self.url_pattern:
- self.definitions['requests_per_url']['lines'].append([elem.description, elem.description,
+ self.definitions['requests_per_url']['lines'].append([elem.description, elem.description[4:],
'incremental'])
self.data.update({elem.description: 0})
- self.data.update({'other_url': 0})
+ self.data.update({'rpu_other': 0})
else:
self.order.remove('requests_per_url')
self._get_data_per_url(match_dict['url'])
# requests per http method
self._get_data_http_method(match_dict['method'])
+ # requests per http version
+ if 'http_version' in match_dict:
+ self._get_data_http_version(match_dict['http_version'])
# bandwidth sent
- self.data['bytes_sent'] += int(match_dict['bytes_sent'] if '-' not in match_dict['bytes_sent'] else 0)
+ bytes_sent = match_dict['bytes_sent'] if '-' not in match_dict['bytes_sent'] else 0
+ self.data['bytes_sent'] += int(bytes_sent)
# request processing time and bandwidth received
if 'resp_length' in match_dict:
self.data['resp_length'] += int(match_dict['resp_length'])
chart_string_copy, 'http_method')
self.data[method] += 1
+ def _get_data_http_version(self, http_version):
+ """
+ :param http_version: str: METHOD from parsed line. Ex.: '1.1', '1.0'
+ :return:
+ Calls add_new_dimension method If the value is found for the first time
+ """
+ http_version_dim_id = http_version.replace('.', '_')
+ if http_version_dim_id not in self.data:
+ chart_string_copy = self.http_version_chart
+ self.http_version_chart = self.add_new_dimension(http_version_dim_id,
+ [http_version_dim_id, http_version, 'incremental'],
+ chart_string_copy, 'http_version')
+ self.data[http_version_dim_id] += 1
+
def _get_data_per_url(self, url):
"""
:param url: str: URL from parsed line
match = True
break
if not match:
- self.data['other_url'] += 1
+ self.data['rpu_other'] += 1
def _get_data_statuses(self, code):
"""
"""
:param pool: list of ip addresses
:param address: ip address
- :param pool_size: current size of pool
- :return: True if address not in pool. False if address in pool
+ :param pool_size: current pool size
+ :return: True if address not in pool. False if address in pool.
"""
index = bisect.bisect_left(pool, address)
if index < pool_size:
else:
bisect.insort_left(pool, address)
return True
+
+
+def find_regex_return(match_dict=None, msg='Generic error message'):
+ """
+ :param match_dict: dict: re.search.groupdict() or None
+ :param msg: str: error description
+ :return: tuple:
+ """
+ return match_dict, msg
+
+
+def check_req_per_url_pattern(string, url_pattern):
+ """
+ :param string: str:
+ :param url_pattern: dict: ex. {'dim1': 'pattern1>', 'dim2': '<pattern2>'}
+ :return: list of named tuples or None:
+ We need to make sure all patterns are valid regular expressions
+ """
+ if not is_dict(url_pattern):
+ return None
+
+ result = list()
+
+ def is_valid_pattern(pattern):
+ """
+ :param pattern: str
+ :return: re.compile(pattern) or None
+ """
+ if not isinstance(pattern, str):
+ return False
+ else:
+ try:
+ compile_pattern = re.compile(pattern)
+ except re.error:
+ return False
+ else:
+ return compile_pattern
+
+ for dimension, regex in url_pattern.items():
+ valid_pattern = is_valid_pattern(regex)
+ if isinstance(dimension, str) and valid_pattern:
+ result.append(NAMED_URL_PATTERN(description='_'.join([string, dimension]), pattern=valid_pattern))
+
+ return result or None
+
+
+def is_dict(obj):
+ """
+ :param obj: dict:
+ :return: True or False
+ obj can be <dict> or <OrderedDict>
+ """
+ try:
+ obj.keys()
+ except AttributeError:
+ return False
+ else:
+ return True
projects / netdata.git / blobdiff