When writing software for the commercial world it is crucial your application can log errors, be it to the database, or to a text file, or even to the Windows Event Log if necessary. The .NET Framework contains a mechanism called a Trace Listener. This Listener is an object which directs the trace output to an appropriate target, such as an error log for example.
In this article I am going to show you how to use the Trace Listener to record your application’s errors to an error log.
To begin this example create a console application and add an App.config file to your project. We are going to use the App.config file to tell our application that it should create a trace listener on start-up. Amend the XML code of your App.config file to look like the below example.
<?xml version="1.0" encoding="utf-8" ?>
<configuration>
<system.diagnostics>
<trace autoflush="true" indentsize="4">
<listeners>
<add name="myListener" type="System.Diagnostics.TextWriterTraceListener" initializeData="application.log" />
<remove name="Default" />
</listeners>
</trace>
</system.diagnostics>
</configuration>
As you can see, we are adding a listener of type System.Diagnostics.TextWriterTraceListener which is set to write to a log file called application.log. Also note that in the <trace> node we are setting the autoflush attribute to true. When set to true this attribute tells the listener to flush the buffer after every write operation. For this example it is great but for real world applications you probably would want to set this to false to limit disk read and write operations.
Now in your static Main method you can write to the log using the Trace class as shown below.
static void Main(string[] args)
{
Trace.WriteLine("Application started.", "MyApp");
Trace.WriteLine("working...", "MyApp");
Trace.WriteLine("Application finished.", "MyApp");
}
If we open the application.log file created automatically by the listener right after we run the above code we will find the following in the file:
MyApp: Application started. MyApp: working... MyApp: Application finished.
As you can see it’s quite easy to write to the log file. We didn’t even have to open a FileStream to write to the file because it was all done for us automatically by the listener.
This logger we just created is quite crude since we are not recording the date and time of the error, nor are we recording the source. So let’s extend the concept a little by creating a small logger class as shown below.
using System;
using System.Diagnostics;
public static class Logger
{
public static void Error(string message, string module)
{
WriteEntry(message, "error", module);
}
public static void Error(Exception ex, string module)
{
WriteEntry(ex.Message, "error", module);
}
public static void Warning(string message, string module)
{
WriteEntry(message, "warning", module);
}
public static void Info(string message, string module)
{
WriteEntry(message, "info", module);
}
private static void WriteEntry(string message, string type, string module)
{
Trace.WriteLine(
string.Format("{0},{1},{2},{3}",
DateTime.Now.ToString("yyyy-MM-dd HH:mm:ss"),
type,
module,
message));
}
}
The above Logger class has four public static methods – Warning, Info, and an overloaded Error method. Each of these methods calls the private WriteEntry method which actually writes to the log through the Trace class. As you can see here we are writing the date and time, the entry type (“error”, “warning”, or “info”), the module where the error occurred, and the actual error message.
Now to use this class we can do the following:
static void Main(string[] args)
{
try
{
Logger.Info("Application started.", "MyApp");
Logger.Info("working...", "MyApp");
Logger.Warning("Application about to exit!", "MyApp");
Logger.Info("Application finished.", "MyApp");
throw new Exception();
}
catch (Exception ex)
{
Logger.Error(ex, "MyApp");
}
}
The above code would produce the following log:
2009-09-16 23:02:04,info,MyApp,Application started. 2009-09-16 23:02:04,info,MyApp,working... 2009-09-16 23:02:04,warning,MyApp,Application about to exit! 2009-09-16 23:02:04,info,MyApp,Application finished. 2009-09-16 23:02:04,error,MyApp,Exception of type 'System.Exception' was thrown.
And there you have it – you now know how to use the Trace Listener in C# and how to create a simple logger class. Obviously the Logger class we created can be enhanced a lot more. For example you could add code to wrap the log file after it grows to 5MB. You could also add code to log the thread which the error occurred on. There is a large number of amendments you could do to make this class better.
I hope you enjoyed this article. Feel free to leave any comments below or if you want you can contact me through my contact page. Also, if you haven’t already subscribed to my rss feed, please do so – Grab RSS Feed.
Dave
My name is David Azzopardi and I'm a software developer by profession. I have been working in the software industry for around eight years now and I have learned a few things through my experiences.
{ 9 comments… read them below or add one }
Perfect example for what I needed. Good and consise explaination too. Thanks!
Glad you found this helpful.
Hi! Great site, great tutorial, it really explained it alle very well… I can only hope for a follow-up on this tracing subject. I would like to have your thoughts about how you would handle the trace log once the software is published…
Thanks again! I will be looking into your site from time to time from now on…
Thanks for your comment. I’m glad you found the tutorial helpful.
I usually handle the trace log in two different ways depending on the application. If the application writes massive amounts of logs to the trace file I would create a new file each day, and then either compress them or archive them every month or so. If the application only writes errors to the log file I would start a new file after it reaches about 5Mb and then archive those every year. It really depends on how much your application will be writing to the log file.
What is your best practise for retrieving the log files? I like to get the log files whin an error occurs so I am sending them to my self by mail from the program…
Usually I do the same. When sending errors by email you will immediately know when something goes wrong, and it’s quite simple to implement and works most of the time.
I have also worked on systems where error logs are automatically downloaded from our clients’ servers over ftp on a weekly basis. Also, what you could do instead of sending emails is open a socket from the client’s pc to your server and send the error message over tcp. Then you could store it in a database on your server, but this does require you to have a server with a static ip always online.
I am sure there are other methods worth exploring, but if you are not expecting many errors and only have a few installations of your application out there, I would stick with receiving errors by email. I would create an email account just for this purpose though, so as not to clutter my personal mailbox.
Thanks, good to hear i am not way of track by mailing myself error/log reports…
(My company is not that big and installations is relatively limited). Logging everything to a database might be a bit to heavy net usage for my liking.
Thanks again for your tutorial and for your comments….
Very good!Very Helpful!
If want to ad date information in log, can add this config:
but ,log will by large.
i want to close trace output when not necessary ,how can i do?
thank you!
Nice article a better way to explain..