So far we love this script.
It is monitoring all of our servers.
I have a problem. We are running tests every 5 minutes on certains objects. (http and ICMP). Here is what I see wrong.
If HM checks the server at lets say 12:00 and its up then its checks it at 12:05 and its down, HM logs 5.00 minutes of down time. What we need is the down time to start the moment the downtime is noted and not for HM to assume the server has been down since its last successfull test.
Please advise.
Alive Time
-
TotalChoice Hosting
- Posts: 1
- Joined: Sun Apr 06, 2003 6:00 pm
- Contact:
>If HM checks the server at lets say 12:00 and its up then its checks it at 12:05 and its down, HM logs 5.00 minutes of down time.
Are you talking about "Alive time", "Dead time", "Alive %", "Dead %" counters? Yes, its how HostMonitor works. So, you think calculate time between 2 consecutive good (or bad) statuses only is more correct way to calculate these counters? But what about situation when test status changes? Do not calculate this time at all? How will look statistics in this case: Total time - 5 hours, Alive time - 4 hours, Dead time - 20 min. Where is 40 more minutes?
I think there is nothing wrong with current method of calculation. Yes, if test status changes to "bad", HostMonitor assumes it was "bad" for whole time interval. But HostMonitor make the same assumption when test status changes back to "good". And this compensate distribution of time between counters.
Regards
Alex
Are you talking about "Alive time", "Dead time", "Alive %", "Dead %" counters? Yes, its how HostMonitor works. So, you think calculate time between 2 consecutive good (or bad) statuses only is more correct way to calculate these counters? But what about situation when test status changes? Do not calculate this time at all? How will look statistics in this case: Total time - 5 hours, Alive time - 4 hours, Dead time - 20 min. Where is 40 more minutes?
I think there is nothing wrong with current method of calculation. Yes, if test status changes to "bad", HostMonitor assumes it was "bad" for whole time interval. But HostMonitor make the same assumption when test status changes back to "good". And this compensate distribution of time between counters.
Regards
Alex