I currently use Host Monitor 6.80 to monitor serveral servers using ping, service, and process tests. When running Log Analyzer for my monthly reports, i noticed that the ping test was returning a 93.33% uptime and the service and process tests returned a 97.42% uptime. This didn't make sense to me and I was wondering if someone could explain how Log Analyzer determines the percentages. Basically, if the server that the process or service runs on is unavailable, does that count against the uptime of the service or will Host Monitor disregard it?
Thanks in advance for any help.
Log Analyzer percentages
There are various options that tells HostMonitor and Log Analyzer how to calculate statistics
1) HostMonitor: Please check options located on Misc Page in the Options dialog ("Reports & Statistics" section)
Display Alive/Dead ratio of passed/failed tests
Display Alive/Dead ratio of alive/dead time
This switch defines how Alive/Dead/Unknown ratio should be calculated: it is either based on quantity of passed/failed probes (HostMonitor version 3 and 4 supported this mode only) or based on alive/dead time (new mode). New mode is useful when you setup HostMonitor to change test intervals for failed tests.
2) Log Analyzer: Take a look on Statistisc page in the Options dialog
100% time = alive + dead + unknown
100% time = alive + dead + unknown + WaitForMaster
100% time = alive + dead + unknown + disabled
100% time = alive + dead + unknown + disabled + WaitForMaster
Regards
Alex
1) HostMonitor: Please check options located on Misc Page in the Options dialog ("Reports & Statistics" section)
Display Alive/Dead ratio of passed/failed tests
Display Alive/Dead ratio of alive/dead time
This switch defines how Alive/Dead/Unknown ratio should be calculated: it is either based on quantity of passed/failed probes (HostMonitor version 3 and 4 supported this mode only) or based on alive/dead time (new mode). New mode is useful when you setup HostMonitor to change test intervals for failed tests.
2) Log Analyzer: Take a look on Statistisc page in the Options dialog
100% time = alive + dead + unknown
100% time = alive + dead + unknown + WaitForMaster
100% time = alive + dead + unknown + disabled
100% time = alive + dead + unknown + disabled + WaitForMaster
"Unavailable"? Are we talking about "Bad" status? Sure, it is used for statistisc calculation. As well as Unknown status.Basically, if the server that the process or service runs on is unavailable, does that count against the uptime of the service or will Host Monitor disregard it
Regards
Alex
-
- Posts: 13
- Joined: Tue Feb 27, 2007 12:06 pm
No, HostMonitor will use Unknown status.Basically, if the server that the service runs on is offline/turned off/etc... will the service or process test return a "failed" status.
Ping and Service tests check the same target server? I assume you are using different test intervals for these items and "Display Alive/Dead ratio of passed/failed tests" statistics option. You may change it to "Display Alive/Dead ratio of alive/dead time"i noticed that the ping test was returning a 93.33% uptime and the service and process tests returned a 97.42% uptime
Regards
Alex
-
- Posts: 13
- Joined: Tue Feb 27, 2007 12:06 pm
-
- Posts: 13
- Joined: Tue Feb 27, 2007 12:06 pm
Sure, Ping test will show more accurate statistics in such case. E.g. network connection can be lost and restored within 40 sec. In such case Ping test will detect the problem while Service test may see nothing because test was performed before connection was lost and after connection was restored.I forgot to say, the test intervals are set to every second for the ping test and every minute for the service test. Will that make a difference?
Regards
Alex
PS
If you have ping test and set of other tests that check the same target server, we recommend to use ping test as master test. In such case HostMonitor will not waste system resources (check services, processes, performance counters) when server does not respond anyway.
http://www.ks-soft.net/hostmon.eng/mfra ... htm#Master
Regards
Alex
If you have ping test and set of other tests that check the same target server, we recommend to use ping test as master test. In such case HostMonitor will not waste system resources (check services, processes, performance counters) when server does not respond anyway.
http://www.ks-soft.net/hostmon.eng/mfra ... htm#Master
Regards
Alex