Tool to monitor HM / RMA

Scooby · Post by **Scooby** » Mon May 18, 2009 8:40 am

I wrote a web application in PHP/MySQL along with a php-based daemon that checks the database and sends the messages.

It is also proving handy to track backup jobs, database dumps and other scheduled tasks to alert me if they failed to execute when expected.

What I wrote is tailored to my environment but it probably wouldn't take much to re-tool it to be more universally applicable.

Paul_NHS · Post by **Paul_NHS** » Mon May 18, 2009 10:09 am

We're an IIS shop, so it probably wouldn't suit.

cheers, Paul

KS-Soft · Post by **KS-Soft** » Sat Jun 27, 2009 12:21 pm

We think HostMonitor can monitor everything including another monitor. However "some people"

still want to have extra application.

Ok, we have included WatchDog application into Enterprise package. http://www.ks-soft.net/hostmon.eng/watchdog/index.htm
Beta version so far...
So we do not need to provide 2nd HostMonitor license for purpose of monitoring HostMonitor anymore.

At the same time we added new test method to HostMonitor - HM Monitor.
http://www.ks-soft.net/hostmon.eng/mfra ... #hmmonitor
It provides much more options than simple TCP check and may monitor various HostMonitor parameters.

Regards
Alex

greyhat64 · Post by **greyhat64** » Tue Jun 30, 2009 11:54 pm

Alex,
You guys really outdid yourself this time!
Really!

KS-Soft · Post by **KS-Soft** » Wed Jul 01, 2009 3:01 pm

I am glad you are happy

Regards
Alex

Robert_in_MTL · Post by **Robert_in_MTL** » Mon Jul 06, 2009 5:26 am

Great News Alex!
Thanks!

meppyman · Post by **meppyman** » Tue Sep 29, 2009 7:48 pm

If I use WatchDog to start a HostMonitor on the WatchDog server the main HM server goes down, how do I cater NT Event Log Tests that are bound by the schedule and last test run time but only if HM and the test is running?

KS-Soft · Post by **KS-Soft** » Tue Sep 29, 2009 8:23 pm

Sorry, I do not understand the question

Regards
Alex

meppyman · Post by **meppyman** » Tue Sep 29, 2009 10:21 pm

I apologise, I wrote that a bit quickly.

It relates to my other post about the problems if HM is restarted or goes down, what happens to the NT Event Log tests for example is inconsistent and easy to misunderstand. I still am unsure how to check what period the test covered in the event logs.

For example:

The primary HM goes down and WatchDog starts a new instance on the WatchDog server. It loads the same tests and log files, etc. Will it act as though it was running the whole time, or will it only check NT Event Logs (for example) from the time the backup HM was started?

If HM was 10 minutes away from running a test the NT Event Log on a 24 hour schedule, I might only get 10 minutes of the NT Event Log looked at for the event instead of the 24 hours. Ideally in a failover situation HM would carry on as if it was always running, so that you have no variation in the tests and alerts.

KS-Soft · Post by **KS-Soft** » Thu Oct 01, 2009 4:20 pm

The primary HM goes down and WatchDog starts a new instance on the WatchDog server. It loads the same tests and log files, etc. Will it act as though it was running the whole time, or will it only check NT Event Logs (for example) from the time the backup HM was started?

It depends on how and when exactly you copy your configuration files from one system to another.
If you just start another instance without any setup, it will not be able to monitor anything. You need to save and copy your configuration files from one system to another. E.g. you may use scheduled HM Script action with SaveTestList and StartProgram commands to save current HostMonitor status (test list) and copy configuration files from one system to another every 10 min.

If HM was 10 minutes away from running a test the NT Event Log on a 24 hour schedule, I might only get 10 minutes of the NT Event Log looked at for the event instead of the 24 hours.

If we are talking about NT Event Log test, then HostMonitor does not check "old" events after restart.
BTW: What is the reason to perform such check just once a day? I think if some problem happens, the sooner you know the better.

Ideally in a failover situation HM would carry on as if it was always running, so that you have no variation in the tests and alerts.

If we are talking about "clustering" monitoring, we have such task in our "to do" list. On the other hand... I am sure there are such software on the marked but as I know such software costs about 50-100 times more that HostMonitor. Sure, if we create such version, we will sell it on lower price, not $100,000 per license. However it will be more expensive than our "regular" Enterprise license. If you need failover monitoring solution on very low price, I think its better to use 2 HostMonitors or HostMonitor+WatchDog and some simple custom made scripts. You may setup such system (with some minor disadvantages) right now.

Regards
Alex

meppyman · Post by **meppyman** » Sat Oct 03, 2009 2:05 am

KS-Soft wrote:You need to save and copy your configuration files from one system to another. E.g. you may use scheduled HM Script action with SaveTestList and StartProgram commands to save current HostMonitor status (test list) and copy configuration files from one system to another every 10 min.

Thanks for the tip, will configure it to do this.

BTW: What is the reason to perform such check just once a day? I think if some problem happens, the sooner you know the better.

Just an example, but in that case the event will only ever be triggered at close to a specific time once a day, so I run the test just after that. I guess this goes with my other post of being able to specify the time range to check the log, in this case I would just check between 5am and 5:20am for example. But the example also applies if you test every 10 minutes, if the system fail-over 30 seconds before the test is about to run then there is a 9-10 min window where the logs will not get checked.

If we are talking about "clustering" monitoring, we have such task in our "to do" list. ... If you need fail-over monitoring solution on very low price, I think its better to use 2 HostMonitors or HostMonitor+WatchDog and some simple custom made scripts. You may setup such system (with some minor disadvantages) right now.

Will be interesting to see that but perhaps it might be priced above the level that is profitable for me with my small business customer monitoring. I think the fail-over will be close to perfect and if the event log workings are tweaked a little to avoid missing events during a failure then this will be good enough for me.

Cheers
Mark