HostMonitor in High Availability

[Prometheus]

Hi,

I know it is possible to setup a second HostMonitor, to monitor the first HostMonitor application. But is it also possible to fully run HostMonitor in HA (High Availability) mode? If so how can I do this?

If it is not in the software yet, can you put it on the wish list?

greets,
Jeffrey

[SplanK]

I reviewed this recently as we had a similar requirement..... who monitors the monitor!

We have 2x sites, one with a HM server.

We use the 'built in' Watchdog application on a 2nd server at the 2nd site to alert us via email if HM goes down after 5/10 minutes (to account for a temp split in the network). Its a simple rule based app that's been quite useful. It launches an external (so I use send mail to send me an alert...), but you could use it to launch a 2nd HM server..

a 2nd HM server would require I believe a 2nd licence, even if only 1 copy will be operational at one time.

[Prometheus]

Hi Splank,

Thank you for the answer, this is really helpfull. Although redundancy is key these days. We currently have HM installed on a virtual machine on site A and a replication of this virtual machine on site B. If A goes down B comes up within 60 seconds. But if the application fails for some reason, im in trouble (we had this a couple of weeks back, application was working but tests did not run due to logging issues. We found out after 3 days). We made some improvements due to this issue and it made us realise we are very dependant of the monitoring.

Bottomline, we need to make sure HM is online and working 24x7 99.9% reliable.

[KS-Soft]

We are working on cluster version.

For now you can use 2nd instance. HM Monitor test method allows you to check various parameters, it could detect logging problem as well.
E.g.
%HMTestsDone% - Number of tests performed by tested HostMonitor (since previous test execution). Returns 0 if request failed
%HMTestsFailed% - Number of "failed' test probes performed by HostMonitor (since previous test execution). Returns 0 if request failed
%HMActionsDone% - Number of actions executed by HostMonitor (since previous test execution). Returns 0 if request failed
%HMLogsDone% - Number of log records added by tested HostMonitor (since previous test execution). Returns 0 if request failed
%HMTestsPerSec% - Average "tests per second" rate for tested HostMonitor since previous test execution
%HMActionsPerSec% - Average "actions per second" rate for tested HostMonitor since previous test execution
%HMLogsPerSec% - Average "log records per second" rate for tested HostMonitor since previous test execution
%HMLoggingPoolUsage% - Current logging pool usage (percentage). High value of this counter means logging performance is too low for your testlist

"Execute HM Script" action can load HML file with tests when you need (LoadTestList command)
http://www.ks-soft.net/hostmon.eng/mframe.htm#actions.htm#actScript

Regards
Alex

[Prometheus]

Hi Alex,

I can work with that, is there any sight on when the cluster option will be available and what it will do with the license?

[KS-Soft]

I think it will be released next year and probably you will need special "cluster" license (upgrade from existing licenses will be available).

Regards
Alex

[Prometheus]

Hi Alex,

Good to hear, can you keep me informed about the progress?

[KS-Soft]

>can you keep me informed about the progress?

Not sure what exactly we should tell you about progress

Regards
Alex

[Prometheus]

If its not to much trouble maybe you can keep me updated when next year the function will be available, and maybe if a beta is up for testing?

[KS-Soft]

What I can say regarding progress - many people want to see "network map with links" function so we changed priority and cluster code will be finished after "the map".

Yes, I think Beta will be available. You will see information and download link on our web site and HostMonitor will inform you when new version will be available (unless you disabled "check for updates" option - Options dialog, Misc page).
So I don't need to do anything special, everything as usual

Regards
Alex