The new Nagios version comes with a bunch of new functionality and fixes. However, this section covers only the most important ones. It is recommended that you view the complete Changelog file that comes with all distributions of Nagios, or the Nagios documentation.

Macro handling is one area where there have been numerous changes. The most important improvement you might notice is that 40 new macros have been added. A notable incompatible change is that the $NOTIFICATIONNUMBER$ macro has been removed in favor of the $HOSTNOTIFICATIONNUMBER$ and $SERVICENOTIFICATIONNUMBER$ macros. Moreover, macros have now been set as environment variables so that your scripts can access them easily. Because this can cause performance issues, there is also an option to disable the environment variables settings of all Nagios macros.

There have been significant changes to how Nagios stores information. The main reason behind this is that Nagios now allows plugins to return multiple lines with information or performance data. This allows more detailed information about hosts and services to be stored. The format of Nagios information and retention files has changed to adapt to this functionality. In the previous format for storing status and many other files, each line was used to represent a single object. This information is now stored in a format similar to Nagios configuration. This requires changes for all applications that read service statuses directly from the Nagios files.

The previous versions of Nagios stored scheduled downtimes as well as host and service comments in separate files. Version 3.x introduced a retention file that stores various information related to hosts and services. This file now contains a list of scheduled downtimes and comments related to each item. It also allows the storage of more information on Nagios restarts, which would be useful when performing frequent restarts.

The embedded Perl interpreter is where a lot has changed under the hood as well. You can now decide whether or not to use embedded Perl in the configuration at compilation time, as it was with 1.x and 2.x. In addition, individual plugins may decide if they want to force enabling or disabling embedded Perl usage. This would mainly be useful if a few of your Perl plugins cause problems with embedded Perl and you don't want to lose the functionality of other plugins that also have embedded Perl.

Now, it is also possible to change the frequency of monitoring hosts or services on the fly. This can be done by sending proper commands to the Nagios daemon. This functionality can be used to modify how often an object should be checked, as well as the time periods during which checks should be performed and notifications should be sent out.

A large improvement has been made in terms of host and service definitions. The major difference is that inheritance can now be done from more than object — this means that your host definition can inherit some attributes from one template and the remaining attributes from another. Services also inherit all contact and notification settings from the host they are running on, unless otherwise specified.

Starting with Nagios 3, it is possible to specify group members by specifying other groups. All objects that are members of specified groups will also be members of this group. For example, when defining a host group, it is possible to specify other host groups. All hosts who are members of such a group will also be members of the currently-described group.

Host checks have also been improved in various ways. Starting with Nagios 3, all checks will be done in parallel, which speeds up Nagios performance enormously. Host check retry handling has also been improved, and now uses the same logic as the service checks.

The dependency system has also been improved in Nagios 3. It is now much easier to configure complex dependencies. This also allows defining dependencies between services on the same host by not specifying dependent host name. It is also possible to set up periods for which dependencies are valid. If nothing is specified, a dependency is valid all the time.

Nagios 3 also introduces support for pre-caching object information. This means that instead of reading all parameters from the configuration file and creating a dependency set for all types of objects, Nagios is able to save its internal representation of the data, so that the next time it starts, it reads the cache instead of re-analyzing the configuration.

Defining time periods also is more powerful now. It is possible to specify date exceptions. For example, combining a definition of Monday to Friday from 9 AM to 5 PM with national holidays will offer a more precise definition of working hours. It is also possible to use a skip date in the time period; for example, 'every 3 days'.

For a complete list of changes, please visit Nagios 3.0 documentation website: http://nagios.sourceforge.net/docs/3_0/whatsnew.html.