I am using log4net in a web app, and log all page errors to a SQL server. I was wondering if there was any way to retrieve the entry ID generated by it. I'm going off of the documentation found here
http://logging.apache.org/log4net/release/config-examples.html
I want to use this ID as a reference number I can show to a customer so that they may contact customer support to lookup in the system and not have to go through a log file.
Apart from writing your own appender as floyddotnet suggested you could consider:
Use a GUID. You can easily generate it in your application and will serve most of your purposes. Drawback: It may be inconvenient for the customers if they try to tell your support stuff about it on the phone. If you have only email support than this is maybe not an issue.
Consider creating an incident number outside of the logging framework. A quick call to a stored procedure that returns an ID that you save in a nullable field in your log table.
A combination of the above: Use a Guid and after logging you call a stored procedure that creates an incident and returns the ID.
Writing an appender that returns the ID creates a dependency between your application and appenders that you normally do not have: Log4net was designed with a clear separation between logging and writing the log messages somewhere. The appender that you need would affect that separation.
Since the ID is generated by the database and not by log4net, I don't believe this information is available to you.
What I've done in using log4net for such conditions is to include a datetime stamp in the message that goes down to the millisecond and present that to the user as a reference number. You can do then do a simple SQL query to get to the message in the log table.
I'm not sure its posible but you can write your own Appender for log4net end store this information in the log4net-context.
Howto writing an appender for log4net:
http://www.alteridem.net/2008/01/10/writing-an-appender-for-log4net/
Context-Description:
http://logging.apache.org/log4net/release/manual/contexts.html
Related
We are pushing logs using "buildfire.publicData.insert". How do we get access to those logs?
Would these logs contain the header that is sent in the calls? We also need to see the source and destination information. Those are the logs we really need.
Public data is really just a data source, think of it as a database where the data you save in you can retrieve later.
So if you're using it to push logs you need to make sure that any needed information is in the body object are passed to save or insert calls. those can be retrieved later using search calls as documented in https://sdk.buildfire.com/docs/public-data
You can also implement the search only on the control side of your plugin if these data are not meant to be shared with widget users.
Since you might have a lot of records it is recommended you use https://sdk.buildfire.com/docs/indexed-fields so your queries are efficient.
I'm setting up logging for Azure service.
Currently, messages I get in wadlogstable look like this:
<Properties>
<EventTickCount SqlType="bigint">635193311660155844</EventTickCount>
<DeploymentId SqlType="nvarchar(max)">deployment21(67) </DeploymentId>
<Role SqlType="nvarchar(max)">HTMLConverterWebRole </Role>
<RoleInstance SqlType="nvarchar(max)">deployment21(67).HTMLConverterWrapper.Cloud.HTMLConverterWebRole_IN_0 </RoleInstance>
<Level SqlType="int">2</Level>
<EventId SqlType="int">0</EventId>
<Pid SqlType="int">6900</Pid>
<Tid SqlType="int">14840</Tid>
<Message SqlType="nvarchar(max)">2013-11-06 12:39:25.8449|ERROR|My error message</Message>
</Properties>
I haven't been to production yet, but I suppose that it's pretty inconvenient to search in xml. What are the best practices for this? Can I customize the elements in it? I don't think I really need Pid, Tid, also I don't see a purpose of EventId.
Update: I'm actually using NLog right now, but I'm doing it as described here: http://awkwardcoder.blogspot.com/2012/03/getting-nlog-working-with-azure-is-as.html
So it posts logs to Trace target and as I understand traces are captured by DiagnosticMonitorTraceListener, ending in Windows Azure table. So I'm using NLog to format my "Message" element in the resulting xml, also "Level" and "EventId" are elements are dependent on which NLog method I call (Logger.Debug* or Logger.Error* etc.), but I don't have access to general format of this xml. Also, I would probably prefer custom logging table with dedicated fields for "Level", "Date" and so on, so I don't have to parse it in each log query.
Unfortunately you don't have control over the format of the data which gets logged automatically by Windows Azure Diagnostics. You could get fine grained control if you use custom logging. For custom logging you could use something like NLog. In that scenario, the data logged by your application is stored in files and get automatically transferred to blob storage using Windows Azure Diagnostics.
You can also use Perfomance Counters + 3rd party tools to display results (i.e New Relic) or you can build your own dashboard.
http://www.windowsazure.com/en-us/develop/net/common-tasks/performance-profiling/
http://www.codeproject.com/Articles/303686/Windows-Azure-Diagnostics-Performance-Counters-In
http://michaelwasham.com/2011/09/19/windows-azure-diagnostics-and-powershell-performance-counters/
I am using Log4Net API in my application to log any important event or information as well as logging my exceptions in database.
There might be an exception while using Log4Net API; if it fails to perform logging into database then how will this exception be stored? Where will all other logging and exception logging be stored in case the Log4Net API fails to perform logging?
Log4net will fail silently so that the primary function of the application is not disturbed. If you have an event that you absolutely need to have in the database, the you should make it part of your business logic i.e. write it yourself to the database.
If you insist using log4net for this, then you could write an appender that has some fail over logic that makes sure that all events get logged to the database (eventually). However this does not seem to be a good idea...
As for normal logging: In order to ensure that I have logs I usually use a rolling file appender in addition to other appenders. This way I can be quite sure that in the worst case I have the log available in a file.
Please help me with this query in using log4net.
I am using log4net in mhy we application. I am facing issues in configuring log4net to log errors at user level.
That is, If user X logs in, I like to create file name X and all error for user X should be written in X.log. Siilarly if Y user logs in the log file should be in name of Y.log and the most important point to note is, they could log in concurrently.
I tried the luck by creating log files whose name would be framed dynamically as soon as the user logs in. But issue here, if they are not using the application at the same time, the log files are creeated with correct name and writing as expected, but if both users have active sessions, log file is created only for user who FIRST logged in and error of second user has been recorded in log file that is created for FIRST user.
Please help me in this.
There has to be a better solution from this one, but you can change log4net configuration from code and even decide which config file to load - so you can do it in code, which is not as nice as editing an XML file.
so what you need to do, which is highly not recommended, is to create log4net configuration each time you call the logger static class, and do what needed based on the calling user.
again.. it doesn't feel right !
(and it will probably perform poorly).
another BETTER solution is to log everything to database (log4net supports it), with a user column, and then produce the logs from db....
We have multiple log files like database log, weblog, quartzlog in our application.
Any log from files under package /app/database will go to database log.
Any log from files under package /app/offline will go to quartzlog log.
What we need now is - want to direct the log staments from one of the java file under /app/database to be outputted to quartzlog instead of database log.
How can we select a particular log file in java file?
You need to define the appropriate appender that logs in the desired file. Read this short introduction to see how you can do it.
Then in the configuration file, you can instruct all messages from a specific package to go in the selected appender:
log4j.logger.my.package = DEBUG, myFileAppender
EDIT:
I believe that in log4j only package resolution is possible - you can't use an appender per file or method. You could try to work around this by adding an extra layer on top of log4j or implementing your own appender.
For example, instead of log.debug use:
my.loggerproxy.log.debug(message);
If you only need to do it from a single method, then the above will be enough. Just instruct the logger proxy package to be logged in a different file.