Most .NET tracing and logging frameworks share a common list of features and concepts.
Advanced logging tools come with additional logging capabilities, filtering mechanism, protocols and
even graphical viewer and analysis tools. Please see below for a list of basic and advanced features.
This list should be a good starting point to learn more about the most important features when
evaluating, comparing and selecting a .NET logging framework.
If you are new to logging and tracing, please see below for a short summary of the key concepts and features
of logging frameworks and libraries.
[Basic] Logging simple text messages
The most basic feature of all logging frameworks is the ability to log and write
simple text messages such as debug information or errors and related information like
timestamps to a log file.
[Basic] Log levels to filter messages
Most logging tools allow developers to specify a log level such as debug, info,
warning and error for each log level and filter these messages by severity. In a production
environment, administrators and support staff might only be interested in warnings and errors,
while developers need more detailed messages during debugging.
[Basic] Multiple connections or targets
Logging frameworks often allow developers to write or send the logging information to multiple
connections or targets and support multiple protocols such as log files or network connections.
The desired connections and targets can be changed at run-time by changing an application setting.
[Basic] Log categories or sessions
Most tracing and logging libraries support a feature to group related logging messages into
categories or sessions. This allows developers to group all log entries of a specific application
module or user session and enable or disable logging for this group of messages independently from other
logging code.
[Basic] Log file rotation
When an application is running for longer periods (such as server applications or
batch processes that are executed daily over multiple months), it is often necessary
to split log files into multiple smaller files. Log file rotation, a feature
supported by most serious logging frameworks, splits log files in chunks of
predefined time periods or file sizes. This could result, for example, in
a separate log file for each day with a unique file name that indicates that day's
date.
[Basic] Loading logging settings from a configuration file
Some logging libraries allow loading logging settings from a configuration file. This allows
developers and support staff to easily change logging settings at run-time without changing
the application or building their own configuration infrastructure.
[Advanced] High logging throughput and performance
As logging is often used during production usage and not just while debugging an application,
performance is a critical factor when choosing a logging framework.
[Advanced] Logging objects and other data
Some logging tools also support logging of complex data structures such as entire objects, binary data,
exception stack traces, screenshots, database results and more, allowing developers to make more sense out of the
logging data compared to just logging simple text messages.
[Advanced] Thread-safe logging
An important feature of serious logging libraries is thread-safety. This means that
it's possible to call and access the same logger object from multiple threads
without causing any deadlocks, race conditions or other thread-related issues. Additionally,
the synchronization mechanism should be as fast and unobtrusive as possible to minimize
thread locking.
[Advanced] Tracing method invocation
Tracing method invocation allows developers to easily understand which methods are called,
in what order they are called and how much time they need to execute. This is especially useful
to optimize and fine-tune code and to understand and follow application logic and errors.
[Advanced] Watching and logging variable values
It's often useful to log and watch variable values separately from log messages in order to
follow variable values and system information such as database connections, thread count and
memory usage.
[Advanced] Logging thread and process behavior
When developing distributed and/or multi-threading applications, it's often useful to know
how many threads and processes are running, what each thread is doing and how long a thread
or process has been running. Advanced logging frameworks allow developers to log thread and
process behavior.
[Advanced] Advanced protocols and options
Advanced protocols such as logging to memory or logging via named pipes are important to
customize logging to specific application requirements. Advanced
protocol options such as reconnect, reconnect interval and buffers allow fine-tuning
of protocols and connections for production usage.
[Advanced] Asynchronous logging and backlog queues
To reduce the blocking of application code in high-performance and scalable
applications, asynchronous logging and backlog queues allow developers to fine-tune
logging in regards to production performance and locking behavior.
[Advanced] Log server application
In situations where a central log storage is needed or when multiple processes should log
to the same file, a log server is needed to receive logging data from multiple sources
and merge them in a central location. Advanced logging frameworks already include
a compatible log server.
[GUI] Graphical viewer application
A graphical viewer application that helps analyzing logging data and monitoring applications
in real-time can drastically reduce the time spent on finding run-time issues, following
application logic and fixing application errors. A graphical viewer application is a great aid
and should be included in any advanced logging package (or should alternatively be available as a
separate download or product).
[GUI] Advanced filtering capabilities
A useful viewer application should allow filtering for different log entry attributes such
as category/session, application, thread ID, process ID, log entry type, timestamp or text message.
Filtering is especially important when analyzing log files with huge amounts of log entries.
[GUI] Navigation functionality
Special features that make it easier navigating through logs such as jumping between log entries,
navigating to methods and finding specific messages is a must-have feature for log viewer
tools.
[GUI] Call stacks, graphs and message details
Separate windows to view call stacks, graphs for variable values and message details
help making sense out of big log files and when monitoring applications in real-time.
[GUI] Multiple views or windows
It's often useful to view log entries of different application modules, threads or processes
in separate views or windows. This allows developers to easily understand what each
thread is working on and where application errors occur. Advanced viewer applications also
allow the automatic creation of views for certain criteria such as for new threads, processes
or application modules.
[GUI] Live monitoring via network or pipes
One of the most useful features of viewer applications is the ability to monitor applications in
real-time by receiving log messages via the network or (named) pipes. This allows developers and
administrators to monitor production systems and spot errors as they occur.
[GUI] Viewing and inspecting application data
Besides viewing log messages and related information such as call stacks, some viewer applications
also render additional application data attached to log entries such as images (e.g. screenshots),
database results, script files, object properties, binary data and more. This makes it easy
to inspect and validate application data.
[GUI] Graphical configuration utility
A graphical utility to create logging configuration files is often useful as the
amount of available logging protocols and options are often quite overwhelming. Such a
tool makes it a lot easier to create configuration files without the need to look up
every protocol option in the online help.
