- StackTrace details take up more space then the message itself, it is difficult to see when an error message starts and ends
- certain messages might appear very often although there are not really indicating an issue (such as "Session expired" message).
Here is Python3 script which reads MapGuide logs files and gets rid of StackTrace details and also filters out uninteresting messages. It adds a line number so that you can find the message in the original log file quickly. It also creates a simple summary for messages and their frequency. As we have two MapGuide servers the script is configured to read two different log file folders.
If you want to use it you need to configure:
- the path for the MapGuide log files
- the path for saving the shortened log file
- error messages you want to exclude
- the number of most recent log files you want to process (such as the 3 most recent ones)
I haven't done much in Python yet - the script doesn't do much error handling and has other shortcomings as well.
Thats how the result might look like:
****************
**************** Logfile: //wsstadt529/logfiles/MapGuide\Error.log
****************
<start>
3 DATE_TIME 2017-03-02 / 01:50:19
4 Success: Server stopped.
<end>
<start>
DATE_TIME 2017-03-02 / 01:51:23
6 Success: Server started.
<end>
<start>
DATE_TIME 2017-03-02 / 07:27:12
444 Error: An exception occurred in FDO component.
445 Error occurred in Feature Source (Library://FS_BEAR/WT_PO_BW_B/Data/TopobaseDefault.FeatureSource): Zeichenfolge ist kein gültiger Filter. (Cause: Zeichenfolge ist nicht korrekt formatiert. , Root Cause: Zeichenfolge ist nicht korrekt formatiert. )
446
<end>
...
<start>
DATE_TIME 2017-03-02 / 13:35:34
953 Error: An exception occurred in FDO component.
954 Error occurred in Feature Source (Library://FS_BEAR/WT_PO_BW_B/Data/TopobaseDefault.FeatureSource): Zeichenfolge ist kein gültiger Filter. (Cause: Zeichenfolge ist nicht korrekt formatiert. , Root Cause: Zeichenfolge ist nicht korrekt formatiert. )
955
<end>
*********************************************
************* Summary ***********************
*********************************************
# 3 :: Error: Failed to stylize layer: ZIM_Anlageobjekte_DynaS_GeoRest_MBR
# 36 :: Error: Failed to stylize layer: ZH_Orthofoto2015
# 1 :: Success: Server started.
# 1 :: Success: Server stopped.
# 27 :: Error: An exception occurred in FDO component.
************* File(s) ***********************
File processed: //wsstadt529/logfiles/MapGuide\Error.log
# of messages:68
# of messages excluded: 51
Here is the script:
import os
import sys
from collections import Counter
""" extract date and time from logfile line """
def extractdatetime(a_text_line):
#DateTime Format in MapGuide-LogFile:
#<2016-12-15T11:00:34> <2015-07-05T12:39:55>
date = a_text_line[1:11]
time = a_text_line[12:20]
return (date, time)
""" checks whether a certain error message should be ignored / excluded from further processing """
def isErrorMessagesToExclude(a_message):
for text_to_find in errormessage_to_exclude:
if text_to_find in a_message:
return True
return False
""" removes the StackTrace details from an error message """
def removeStackTrace(a_message):
pos = a_message.find('StackTrace:')
if pos > -1:
return a_message[0:pos]
else:
return a_message
""" saves file """
def saveFile(text, filename, mode):
f = open(filename, mode)
f.write(text)
f.close()
""" returns a sorted list of files for a given directory , sorted by date
see: http://stackoverflow.com/questions/4500564/directory-listing-based-on-time
"""
def sorted_ls(path):
mtime = lambda f: os.stat(os.path.join(path, f)).st_mtime
result = list(sorted(os.listdir(path), key=mtime))
return result
""" keeps only MapGuide Logfiles and filters out all other files """
def filterLogfiles(files_in_dir):
filtered_files = []
for filename in files_in_dir:
if filename.endswith(".log") and filename.startswith("Error"):
filtered_files.append(filename)
return filtered_files
""" processes a single MG log file and simplifies content """
def processLogFile(logfile):
# counter for line number in log file
line_number = 0
# counter for number of messages processed
counter_messages = 0
# counter for messages we exclude from processing
counter_messages_ingnored = 0
# internal counter
count_opening_tag = 0
message = ''
message_part = ''
# encoding utf-8, ascii : returns error message while reading a certain character/bytes
with open(logfile, encoding='“latin-1') as a_file:
for a_line in a_file:
line_number += 1
# all messages start with '<' and first line also contains date&time
if '<' in a_line:
count_opening_tag +=1
#get date and time
str_date, str_time = extractdatetime(a_line)
a_line = 'DATE_TIME ' + str_date + ' / ' + str_time + '\n'
# we processing the first line of the current message
if count_opening_tag == 1:
# line number and date/time information in one new line
message_part += str(line_number) + ' ' + a_line
# we have reached the first line of the following message - now we need tp process the previous message
if count_opening_tag == 2:
#first we check whether the message can be ignored
if isErrorMessagesToExclude(message_part) is False:
# we remove the StackTrace details
message_part = removeStackTrace(message_part)
counter_messages += 1
# we wrap the message text in <start><end> tags
message += '\n<start>\n'
# we add the processed message to the result
message += message_part
message += '\n<end>'
else:
counter_messages_ingnored += 1
#as this is the first line of the "next" message already a_line contains the DATE_TIME for it
message_part = a_line
# reset counter - the current line is the first line of the next message
count_opening_tag = 1
# last line reached - last message block is not yet fully processed
# code from above is repeated here to close the processing of last message in logfile
if isErrorMessagesToExclude(message_part) is False:
message_part = removeStackTrace(message_part)
message += '\n<start>\n'
message += message_part
message += '\n<end>'
counter_messages += 1
else:
counter_messages_ingnored += 1
temp = ["File processed: "+logfile, "# of messages:"+str(counter_messages), "# of messages excluded:\t "+str(counter_messages_ingnored) ]
summary_files.append(temp)
print("File processed: "+logfile)
print("# of messages:"+str(counter_messages))
print("# of messages excluded:\t "+str(counter_messages_ingnored))
return message
""" converts the newly created log file(s) summary into a list """
def convertToList(processedlogfile):
with open(processedlogfile, encoding='“latin-1') as a_file:
error_message = False
list_final = []
date_temp =''
time_temp = ''
error1_temp = ''
error2_temp = ''
line_counter = 0
for a_line in a_file:
line_counter += 1
if '<start>' in a_line:
line_counter = 1
error_message = True
if '<end>' in a_line:
error_message = False
line_counter = 0
list_temp = [date_temp, time_temp, error1_temp, error2_temp]
list_final.append(list_temp)
if error_message:
if line_counter == 2:
date_temp = a_line[10:20]
time_temp = a_line[23:]
if line_counter == 3:
error1_temp = a_line[5:].strip()
if line_counter == 4:
error2_temp = a_line[5:].strip()
return list_final
""" iterates over all relevant files and creates summary"""
def processLogfiles(logfile_dir, saveLogFileName, number_of_most_recent_files):
# create a new file for the results
saveFile('', saveLogFileName, 'w')
# get all files from directory with MapGuide logs
files_in_directory = sorted_ls(logfile_dir)
# filter out all non MapGuide log files
files_in_directory = filterLogfiles(files_in_directory)
# only process the most recent files
if(number_of_most_recent_files > 0):
index = ((number_of_most_recent_files ) * -1)
files_in_directory = files_in_directory[index:]
# interate over relevant log files
for filename in files_in_directory:
fn = os.path.join(logfile_dir, filename)
# process single log file
log_file_short = processLogFile(fn)
# create header
header = "\n\n****************"
header += "\n**************** Logfile: " + fn
header += "\n****************\n\n"
# write to file
saveFile(header, saveLogFileName, 'a')
saveFile(log_file_short, saveLogFileName, 'a')
# all files have been processed and relevant information has been written into one file
# now we want to get a summary of logged issues
resultList = convertToList(saveLogFileName)
'''
resultList is list of lists, lists have 4 items each
items 3 and 4 are equal to line 1 and 2 of a message
now we just count item 3 and get a Dictionary where frequency of message is key and message itself is value
'''
res = (Counter(mysublist[2] for mysublist in resultList))
text = "\n\n*********************************************"
text += "\n************* Summary ***********************"
text += "\n*********************************************\n\n"
for message, number in res.items():
text += "# " + str(number)+ "\t :: " + message +'\n'
# append summary
saveFile(text, saveLogFileName, 'a')
summ = "\n************* File(s) ***********************\n"
for alist in summary_files:
summ += "\n".join(alist)+"\n"
# append summary
saveFile(summ, saveLogFileName, 'a')
if __name__ == ("__main__"):
""" add any message you want to ignore when processing the log files"""
errormessage_to_exclude = [
'Session has expired',
'Resource was not found: Session:',
'Die Sitzung (',
'Error: Authentication failed'
]
# number of most recent log files to process
# 0 - for all files to be processed
# can be overwritten when python script is called with parameter
number_most_recent_files = 3
#if argument is provided we assume its a number
if len (sys.argv) == 2 :
number_most_recent_files = int(sys.argv[1])
# path to MapGuide error log directory
logfile_dir1 = '//wsstadt529/logfiles/MapGuide'
logfile_dir2 = '//wsstadt516/logfiles/MapGuide'
# file name for result - simplified log file:
saveLogFileName1 = 'c:/temp/aims_log_processed_529.txt'
saveLogFileName2 = 'c:/temp/aims_log_processed_516.txt'
# start processing
summary_files = []
processLogfiles(logfile_dir1, saveLogFileName1, number_most_recent_files)
summary_files = []
processLogfiles(logfile_dir2, saveLogFileName2, number_most_recent_files)
# open file(s) in Editor
os.startfile(saveLogFileName1)
os.startfile(saveLogFileName2)
No comments:
Post a Comment