Much more flexible alert profiles
Alert profiles may now contain a number of alert actions that can be launched in a predefined order depending on the test results.

Example 1: You want a fresh network stats report to be posted on your intranet web server every time the test finds the server to be up and running. This can be done simply by adding the following action to the alert profile:
  Action: Generate reports
  Start when: 1 consecutive "good" result occur
  Retries: until status changes

Example 2: When some critical service dies, you want the server to automatically reboot. If that does not help, an e-mail should be sent to the on-call technician. If, however, the server remains silent during the next three tests, the network administrator is to be paged until the server is brought back up. To implement this behaviour, create an alert profile with three actions defined like these:

1. Action: remote reboot
   Start when: 1 consecutive "bad" result occur
   Retries: 1
   
2. Action: send e-mail
   Start when: 2 consecutive "bad" results occur
   Retries: 1
   
3. Action: send message to pager
   Start when: 5 consecutive "bad" results occur
   Retries: until status changes
   

Schedules can now be applied to actions the same way they were used for tests. With schedules, actions can be customized per time of the day. For instance, an action profile can be set up to page both the IT manager and the network administrator during regular office hours, and to page the administrator alone the rest of the time, while doing nothing else but writing to the log on weekends. To make this possible, simply check the "Time restriction" option and select an appropriate schedule.

Folders
New to version 3.0, folders bring a new level of control and flexibility to network monitoring. Learn how to organize tests in a tree-like structure (pretty much like you organize files on your hard drive) and take advantage of the property inheritance mechanism.

The folder tree is displayed in the left-hand panel of the HostMonitor main window. Appearing on the right side, in the list view area, are tests contained in the current (selected) folder and (if the appropriate option is selected) its descendent subfolders. Sure enough, all tests are performed based strictly on their schedules, regardles of whether they are currently displayed or not.

Each folder may contain tests and/or nested subfolders, and has a number of properties:

• visible fields - a list of test attributes to display;
• sort order - a list of attributes to sort by;
• Color palette - a customizable color scheme;
• reports - up to 6 individual reports can be associated with a folder.

A new folder automatically inherits properties defined for the parent folder (the top folder inherits the common settings defined in the Options dialog). Once a folder is created, you can change its properties as needed, rename or delete the folder, create tests and nested subfolders, and so on. HostMonitor supports Drag & Drop operations, allowing you to easily move folders around, change their display order, and move tests from one catalog to another.

For example. you can set up a folder representing your company, define a set of report profiles (for the administrator, IT manager, technical support crew, etc.), then create several subfolders, one for each of the company's offices, and you are done! HostMonitor will be automatically generating uniform reports for all company offices (each report will contain test results specific to the office).

In another example, folders can be created based on a test type (Ping, CPU Usage, etc.). For each folder, define a different set of visible columns, depending on the type of test. Thus, for a ping test folder a logical choice would include the "Reply" field, whereas for a CPU Usage folder catalog the "Average reply" field might make sense.

Color profiles
Fully customizable color schemes are available now for a variety of program windows, HTML log files and reports. Color schemes can be assigned to either a catalog or a report type. For example, a report designed for the IT manager might have an entirely different look and feel as compared to the one intended for use by the network administrator. Similarly, reports designed for the field offices may differ in palette from those used by the headquarters.

The following elements have customizable color attributes (both fore- and background):
- underlying layer;
- table header (HTML reports only);
- folders;
- system messages (HTML logs only);
- test colors are status-dependent:
  - "Not testes yet";
  - "Good";
  - "Bad";
  - "Unknown";
  - "Test in progress";
  - "Disabled";
  - "Waiting for master";
  - "Out of schedule ".

Mail profiles
Multiple message templates can now be created for e-mail, pager and ICQ notification. For instance, you can set up HostMonitor to send general server status emails to the manager, and fire off an elaborate problem report to the administrator in case of a network failure.
As before, macro variables are supported in message templates to be substituted with their actual values at message generation time. For a list of the new macros introduced in version 3.0 please see new variables.
The other new feature allows you to specify a particular character set for your templates. This comes in especially handy if you work in a multilingual environment.

Report profiles
The highly flexible Report Manager allows you to create and customize reports to your liking in a variety of ways.
The configurable report parameters include:
- report type: HTML, WLM, DBF or text;
- field list (e. g., test name, status, average response time, downtime, etc.);
- test type filter (e. g., exclude any ping tests);
- status filter (e. g., only include tests with a "Host is alive" or "Unknown" status);
- color scheme (for HTML reports only);
- HTML header and/or footer file (for HTML reports only).

Below is a list of the macro variables that can be used in an external header/footer:

%Date%	-	current date
%Time%	-	current time
%DateTime%	-	current date & time
%BGColor%	-	background color
%TextColor%	-	text color
%LinkColor%	-	hypertext link color
%VLinkColor%	-	visited link color
%ALinkColor%	-	active link color

It is also worth noting that each folder may contain its own list of reports, and each of the reports can be set up with a launching schedule specific to that folder.

• "Unknown" has a new meaning: the program is unable to retrieve the necessary information (e. g., cannot retrieve CPU usage data from a remote machine). What used to be tagged as "Unknown" now has a status of "Not tested yet".
• "Waiting for master" means the test is defined as dependent, i. e. based on another (master) test's results, and the current status of the master test precludes the dependent one from being launched.
• "Out of schedule" - tests are marked with this status if the time to launch the test falls into a disabled time frame according to the schedule specified for that test.

New option: "Treat Unknown status as Bad"
The new "Treat Unknown status as Bad" option can be applied to tests. With this option enabled, if test results cannot be obtained, actions are triggered by HostMonitor the same way as if the test returned a "Bad" status.

The option is not applicable to some of the test types. Please refer to the table below for the meaning of "Unknown" status for diferent test types:

Test	Option applicable	"Unknown" status condition
Ping test	Yes	Host name cannot be resolved into IP address
TCP test	Yes	Host name cannot be resolved into IP address
URL request	No
Disk free space	Yes	Disk drive information is unavailable (e. g., the drive is unmapped)
Folder/File size	Yes	File/folder doesn't exist
File exists	No
Compare files	No
File CRC	Yes	File doesn't exist or read error
UNC	Yes	Resource is unavailable (option is applicable for free space check only)
Oracle	No
Interbase	No
Ms SQL	No
My SQL	No
Postgre SQL	No
Sybase	No
Process	Yes	Information (list of processes) cannot be retrieved
Service	Yes	Connection to CS Manager cannot be established
SNMP	Yes	Information cannot be retrieved
NT Events Log	Yes	Information cannot be retrieved
CPU Usage	Yes	Information cannot be retrieved
External test	Yes	HostMonitor cannot execute external program (program doesn't exist) or program doesn't return ERRORLEVEL within a specified time
ODBC Query	Yes	SQL query fails, or the data field specified is not found in the result set

Please note: the HostMonitor icon displayed in the System Tray may have one of the three colors:
- blue - all tests performed have been successful;
- yellow - "Unknown" status tests have occured;
- red - "Bad" status tests have occured.

New option: "Reverse alert"
When this option is selected, HostMonitor regards test results that are normaly considered benign, as abnormal ones, and vice versa.
This can be useful in some situations. For example, CPU Usage test returns a "Ok" status if the processor load ratio is below a predefined value. With the "Reverse alert" option turned on, the test will work the other way around, so the administrator will be alerted when the processor load is below a predefined value.

Here is how the conversion works for different status values:

Original status	Reverse status
Alive	Bad
Dead	Ok
Ok	Bad
Bad	Ok
BadContents	Ok

Other status values are not affected.

New test property: "Private log"
As before, HostMonitors records test data in a single log file shared by all tests. Additionally, for each test you can now define its own, "private" log file. The type of a private log is determined based on the file extension: HTML for .htm*, DBF for .dbf, plain text for any other extension.
Private logs can be configured to work in one of the following modes:

• Brief - record information on a changed status;
• Full - record all test results, regardless of whether the status changes or not;
• Default - use the common settings defined in the Options dialog.

New test property: "Related URL"
For each test, a related URL can be specified. By adding this property to the list of visible columns or to a report, you will make the related URLs appear in the test list view or report sheet. In HTML reports, URLs appear as hyperlinks. In the list view, tests with a related URL will have a popup menu item that launches the associated program (typically, your default Internet browser).

New test property: "Irregular schedule"
As before, tests in version 3.0 can be configured to execute in regular time intervals, for instance, every 10 minutes. In addition, the new property, "Irregular schedule", allows you to have your tests run on a given day of week or month, at the time specified, like at 8:30 AM every Monday, or at 10:00 PM on the 1st of each month.

Dependencies
For each test, you can now define a list of other tests it depends on, or Master tests. Choose the condition to perform the dependent test based on the status of a Master test:
- "Alive";
- "Alive" or "Unknown";
- "Dead";
- "Dead" or "Unknown".
As an example, consider a network configuration where the accessibility of the company's several web servers depends on the state of a single router. When the router is down, there is probably little sense in trying to do the dozens of tests designed to collect information on the web servers.
That was, of course, a very simplified example. With this new feature, you can implement quite sophisticated dependencies involving a number of master tests which, in turn, depend on other tests' results. For more information on dependencies, please, see the following options: "Syncronize counters", "Syncronize alerts", "Consider status of the master test obsolete after N seconds".

New option: "Synchronize counters"
This option only applies to tests that have one or more master tests. When the option is turned off, and some test is not being launched because its launch condition has not been met, HostMonitors simply marks such a test with the "Wait for Master" status. If, however, the option is turned on, the dependent test receives the status of the master test, and its statistics information is accordingly updated. Thus, if a router on which other tests depend has been tested to a "No answer" status, the same status will be assigned to all dependent tests with the "Synchronize counters" option on, which in turn will increment their respective counters like "Dead time", "Failed tests", etc.

New option: "Synchronize alerts"
This options is applicable only if "Synchronize counters" is also selected. When the option is turned on, besides synchronizing counters, HostMonitor also does action dispatching based on the status propagated from the master test.

Example: On your corporate network you have 10 web servers and 1 router that is critical to the entire network. A router Ping test is set up to page the administrator if the router does not respond. So when the router goes down, and the "Synchronize alerts" option is turned off, HostMonitor sends a message to the administrator's pager and suspends all dependent tests. If "Synchronize alerts" is on, in addition to the router problem notification, actions defined on the dependent tests will be fired off as well.

Improved tests:

• Ping test: A packet loss threshold can be specified for a Ping test, based on which the test status is determined. That is, in the previous versions of HostMonitor a test was considered failed if no echo packets were received (100% loss). Now you can set the option, "Status is bad when N (or more) % of packet lost".
  The other new feater is, the "Reply" field can be set up to represent either the reply time, percentage of lost packets, or percentage of received packets.
• External test: A window state of the program to launch can be specified (e. g., Normal, Minimized, Maximized, Hidden, Non-active, etc.)
  

New test: ODBC query
This test allows you to check the availability of an ODBC data source, run an SQL query, and analyze the value of a specified data field in the result set returned. To set up a test, simply pick one of the ODBC data sources defined in your system, and fill in the Login, Password and Timeout fields. Optionally, provide an SQL query and a condition to check. If no condition is specified, the two possible test status values are "Alive" and "No answer". If a condition is provided, test status can be either:
- "Ok" - if the condition is met;
- "Bad" - if the condition is not met;
- "Unknown" - if the query fails, or the data field specified is not found in the result set.

Advanced logging
As was mentioned above, in addition to the common log, test-specific private logs can now be defined. Another new option allows you to log test information into an ODBC data source. The query to logging test data is entirely up to you. Use the macro variables available in version 3.0 to put in as much detail as you want. For instance, your log SQL command might look like this:
INSERT INTO HMLOG (DATE1, TIME1, TEST_NAME, STATUS, REPLY) VALUES (%Date%, '%Time%', '%TestName%', '%Status%', '%Reply%')
Note: An appropriate database table has to be created before you start using the ODBC logging.

Also new in version 3.0 is the ability to separate system messages (generated by HostMonitor) from test status records. To enable this, set the "System log: use different log file XXXXXX" option.
The settings for these new options are located on the new page, "Advanced options", in the Options dialog.

• Show tests in subfolders (recursive)
• List of visible columns and their order
• List of sorted columns

• Store statistics in HML file - this option tells HostMonitor to write all statistical information into a file, so that the statistics build up across multiple program sessions. If the oprion is turned off, when an HML file closes down, the statistics pertaining to its tests get reset. Please note that statistics for any or all of the tests can be nullified at any time using the "Reset ..." menu command.

• Store "Recurrences" in HML file - this is an important option, because it's the "Recurrences" field that the programs looks at to determine what actions and when to fire off. Therefore, if the option is deselected, upon closing an HML file or shutting down HostMonitor, the counter will be reset, so that when the program is restarted, actions will be triggered as tests are first performormed. With the option turned on, recurrence counters will be saved on shutdown, and at the next startup the program will pick up where it left off, as if the program execution was not interrupted.

• Reset "Recurrences" when test gets disabled
• Remember current folder

• Consider status of the master test obsolete after N seconds - this parameter is used by the program to determine whether the Master test status is up to date. Before starting a dependent test, HostMonitor checks the status values of the Master tests defined on it. The parameter mandates for how long a Master test status is considered to be current.

• Startup delay - this parameter is only applicable if HostMonitor is started as a Win32 service. It comes in handy if you want to monitor some other services that might start after HostMonitor does. With this option turned on, the actual monitoring is deferred at HostMonitor startup for the amount of time specified.

• When service is being stopped: save changes in HML file; discard all changes
• Authentication method (Plain, Login, MD5, or POP before SMTP), login and password for primary and backup SMTP server
• Record info about successful actions
• Record info about failed actions

Group test operations
Test operations (Delete, Enable, Disable, Refresh) can now be performed on a group of selected tests. To select multiple tests use key combinations: Shift + Arrow key for a group of contiguous tests, Ctrl + Arrow key for arbitrary tests.
Additionally, some common test parameters can be edited for a group of selected tests: alert profile, schedule, test interval, comment, related url, private log, etc.

Creating multiple tests in one shot
A range of IP addresses rather than a single adress can be specified for a Ping or TCP test (e. g., 192.52.233.1 - 192.52.233.250). For each of the range addresses, HostMonitor will create a separate test.

Encrypted HML and LST files
Because some test settings may include a password, HostMonitor configuration files are now stored in an encrypted form. For security reasons, it is also recommended to have the "Protect all settings by password" option turned on (located on the Miscellaneous page in the Options dialog).

New linking mechanism for schedules and profiles
You can now rename schedules, alert profiles and other profiles without having to worry that the link between the test and the profile will be broken. The new linking mechanism implemented in version 2.0 uses internal IDs that are unique and non-changeable throughout the lifetime of an object.