1) In reference to the Text Log test, it would be nice to have the option to define a test expression for a pass condition. When monitoring a text log file that contains both online and offline state change type messages, there currently exists no way to enter a test expression that would check for a pass message within the confines of the same Text Log test. Also, adding some additional processing features to this test such as those offered by the Active Script feature would help to make this test more versatile.
2) Create a new Active Script action for use in Action Profiles.
3) Add the ability to assign an optional test execution sequence number to each test. Then for tests contained within the same folder, and with the folder Specials Property "Non-simultaneously test execution" selected, tests would execute in the defined order.
My wishlist...
-
KS-Soft Europe
- Posts: 2832
- Joined: Tue May 16, 2006 4:41 am
- Contact:
Re: My wishlist...
Could you provide wider explanation of your thoughts? Do you mean you want to use some regular expression within "Text Log" test method? Or do you want to set the final status depends on apearing of several strings in the log? Probably, you may use the solution, described in the following topic: http://www.ks-soft.net/cgi-bin/phpBB/vi ... 39&start=0JimMc wrote:1) In reference to the Text Log test, it would be nice to have the option to define a test expression for a pass condition. When monitoring a text log file that contains both online and offline state change type messages, there currently exists no way to enter a test expression that would check for a pass message within the confines of the same Text Log test. Also, adding some additional processing features to this test such as those offered by the Active Script feature would help to make this test more versatile.
Does it suit your needs?
Actually, you may easily implement this by using "Execute external program" action. You just should specify command line string like: cmd /c cscript some_script.vbsJimMc wrote:2) Create a new Active Script action for use in Action Profiles.
Hm. You know, each test depends on its own schedule, so it would be hard to adjust sequence number with shedule options in some cases. We should think a bit about it....JimMc wrote:3) Add the ability to assign an optional test execution sequence number to each test. Then for tests contained within the same folder, and with the folder Specials Property "Non-simultaneously test execution" selected, tests would execute in the defined order.
Regards,
Max
Hi Max,
Before I begin I would simply like to preface the following responses with the comment that I did submit my suggestions to the Wish List area of the Host Monitor forum. I acknowledge that to achieve many of the results I seek it is possible to write scripts or batch files and execute them from Host Monitor. However if an existing feature can be improved by adding an enhancement, that certainly would prevent a lot of coding and debugging effort for many users with similar requirements.
My intent was only to make a need known to the HM developers. If enough users request a similar enhancement or if the HM developers feel an enhancement is worthy of inclusion, then my request posted in this forum was worth submitting.
In reference to request #1
I have an application that communicates over the internet with many pieces of remote equipment. The application identifies each piece of equipment with a unique name. Using these unique names for identification, the application will write status messages into a single log file when equipment status events occur.
Equipment events written to the log file will typically indicate an offline/online condition. For example, a message will be written indicating when a piece of equipment goes offline and another message will be written when the equipment returns to an online state. These events may be very brief (lasting just a few seconds) or in severe cases a piece of equipment may go offline for hours before returning to an online state.
I would like to use the Text Log test to monitor this log file. However the current implementation does not allow the definition of a Look For pass expression only a Look for fail expression. This has several drawbacks.
1) Should a piece of equipment go offline (Fail) for a period greater than the test interval, the Text Log test will return to a Pass state even though the equipment remains offline. This occurs because the test will not find a message in the log file that matches the Fail expression.
2) Should a piece of equipment go offline (Fail) for brief period and then return online, the Text Log test will return a Fail state even though the equipment is online. A Pass condition will only be reported when the next test interval occurs and no match is found for the Fail expression.
3) Between Text Log test intervals, if the piece of equipment goes offline for just an instance and returns to an online state, or if it goes offline and online for several brief instances, I really don't want to be notified about each one of these brief events. I just want to be notified if the cumulative event state between test intervals has changed from that of the previous test. For example...
Previous Test Result --- Current Test Result --- Action
Passed --- Passed --- Do nothing
Passed --- Failed --- Fail Action
Failed --- Passed --- Pass Action
Failed --- Failed --- Do nothing
(Sorry I do not know how to create columns)
The best solution I found was to define a Boolean test expression consisting of both offline and online strings joined by a logical or. I set the test to warn of all new events and set the action profile to repeat until status changes. This type of test is defined for each piece of equipment using the unique name in the expression. This prevents ambiguous test results from occurring if instead I were to use a single Text Log test.
Of course the problem with this approach is that I am warned every time a brief equipment event occurs. Another minor problem is that HM will log online state changes as failures. This is really only an issue if I were concerned about pass/fail statistics.
Certainly it is possible to use some form of external routine or write a vb script to do what I want. However the Text Log test does most of what needs to be done. It just needs a little tweaking to make it a bit more powerful.
In reference to request #2
Yes, I am aware of wscript and cscript and how to execute vb scripts from the command line. What I really like about the Action Script feature is that I have the option to edit and store the script within the HML file. This may not seem like much, but I like having HM keep track of things rather than me chasing external files around the hard drive.
Web pages drive me crazy keeping track of css files, and include files, and graphics files, and javascript files, and on and on and on. I just want HM to be a sanctuary where the application takes care of all the housekeeping tasks for me. It would be nice to have this feature as an available option in the Action Profile.
In reference to request #3
In such instances maybe a folder interval property could override the test interval property for any tests defined within a folder. The reason I added this request to my wish list is in reference to a question I submitted to tech support. I execute a master test that will test my HM server internet connection. If the master test passes I want to execute a number of subordinate tcp connection tests to different ip addresses. If my HM server momentarily looses the internet connection after the master test has passed, a bunch of subordinate test failures will occur.
What I am seeking is to force a master test just prior to executing the subordinate tests. The only way I can come close to achieving this is by adjusting the "Options -> Behavior -> Consider the master test obsolete after" property to a rather small value. Doing so has several undesirable effects...
1) This property is global to the application so it will affect all master tests.
2) Depending on the number of subordinate tests, it can increase test loading especially when using a test such as the tcp test where the test timeout might be lengthy. As a result, the master test could be re-executed many times.
3) It is possible that the subordinate tests may not execute immediately after the master test was completed and the state of the internet connection may have actually changed.
In reference to this situation, due to the dynamics of the internet I realize it is very difficult to create an absolutely effective master/subordinate test scenario. However by allowing more granular control over the testing order it can help to minimize these types of problems.
Thanks,
Jim
Before I begin I would simply like to preface the following responses with the comment that I did submit my suggestions to the Wish List area of the Host Monitor forum. I acknowledge that to achieve many of the results I seek it is possible to write scripts or batch files and execute them from Host Monitor. However if an existing feature can be improved by adding an enhancement, that certainly would prevent a lot of coding and debugging effort for many users with similar requirements.
My intent was only to make a need known to the HM developers. If enough users request a similar enhancement or if the HM developers feel an enhancement is worthy of inclusion, then my request posted in this forum was worth submitting.
In reference to request #1
I have an application that communicates over the internet with many pieces of remote equipment. The application identifies each piece of equipment with a unique name. Using these unique names for identification, the application will write status messages into a single log file when equipment status events occur.
Equipment events written to the log file will typically indicate an offline/online condition. For example, a message will be written indicating when a piece of equipment goes offline and another message will be written when the equipment returns to an online state. These events may be very brief (lasting just a few seconds) or in severe cases a piece of equipment may go offline for hours before returning to an online state.
I would like to use the Text Log test to monitor this log file. However the current implementation does not allow the definition of a Look For pass expression only a Look for fail expression. This has several drawbacks.
1) Should a piece of equipment go offline (Fail) for a period greater than the test interval, the Text Log test will return to a Pass state even though the equipment remains offline. This occurs because the test will not find a message in the log file that matches the Fail expression.
2) Should a piece of equipment go offline (Fail) for brief period and then return online, the Text Log test will return a Fail state even though the equipment is online. A Pass condition will only be reported when the next test interval occurs and no match is found for the Fail expression.
3) Between Text Log test intervals, if the piece of equipment goes offline for just an instance and returns to an online state, or if it goes offline and online for several brief instances, I really don't want to be notified about each one of these brief events. I just want to be notified if the cumulative event state between test intervals has changed from that of the previous test. For example...
Previous Test Result --- Current Test Result --- Action
Passed --- Passed --- Do nothing
Passed --- Failed --- Fail Action
Failed --- Passed --- Pass Action
Failed --- Failed --- Do nothing
(Sorry I do not know how to create columns)
The best solution I found was to define a Boolean test expression consisting of both offline and online strings joined by a logical or. I set the test to warn of all new events and set the action profile to repeat until status changes. This type of test is defined for each piece of equipment using the unique name in the expression. This prevents ambiguous test results from occurring if instead I were to use a single Text Log test.
Of course the problem with this approach is that I am warned every time a brief equipment event occurs. Another minor problem is that HM will log online state changes as failures. This is really only an issue if I were concerned about pass/fail statistics.
Certainly it is possible to use some form of external routine or write a vb script to do what I want. However the Text Log test does most of what needs to be done. It just needs a little tweaking to make it a bit more powerful.
In reference to request #2
Yes, I am aware of wscript and cscript and how to execute vb scripts from the command line. What I really like about the Action Script feature is that I have the option to edit and store the script within the HML file. This may not seem like much, but I like having HM keep track of things rather than me chasing external files around the hard drive.
Web pages drive me crazy keeping track of css files, and include files, and graphics files, and javascript files, and on and on and on. I just want HM to be a sanctuary where the application takes care of all the housekeeping tasks for me. It would be nice to have this feature as an available option in the Action Profile.
In reference to request #3
In such instances maybe a folder interval property could override the test interval property for any tests defined within a folder. The reason I added this request to my wish list is in reference to a question I submitted to tech support. I execute a master test that will test my HM server internet connection. If the master test passes I want to execute a number of subordinate tcp connection tests to different ip addresses. If my HM server momentarily looses the internet connection after the master test has passed, a bunch of subordinate test failures will occur.
What I am seeking is to force a master test just prior to executing the subordinate tests. The only way I can come close to achieving this is by adjusting the "Options -> Behavior -> Consider the master test obsolete after" property to a rather small value. Doing so has several undesirable effects...
1) This property is global to the application so it will affect all master tests.
2) Depending on the number of subordinate tests, it can increase test loading especially when using a test such as the tcp test where the test timeout might be lengthy. As a result, the master test could be re-executed many times.
3) It is possible that the subordinate tests may not execute immediately after the master test was completed and the state of the internet connection may have actually changed.
In reference to this situation, due to the dynamics of the internet I realize it is very difficult to create an absolutely effective master/subordinate test scenario. However by allowing more granular control over the testing order it can help to minimize these types of problems.
Thanks,
Jim