I have had several instances now where lots of alerts coming in rapidly (over a 100 within a short period), causes Host Monitor to either stop monitoring or if it does continue monitoring it stops sending emails and reports. I think I solved the problem of the complete stopping of monitoring by eliminating screen popup alerts. But I still have the problem of the email and reports stopping. I have solved this so far by stopping and restarting the Advanced Host Monitor service, then the emails and reports resume. I am using Windows 2000 Server to run Host Monitor and almost all tests are executed remotely by RMA.
Do you have any suggestions for improving this? I plan to move to Windows 2003 Server very soon, so maybe that OS will be more stable.
Thank you.
Too many alerts jam up Host Monitor
Could you please provide more information:
1) Version of HostMonitor?
2) What kind of alerts do you use? "Send e-mail", "Generate reports" and some other?
3) Do you need to start all these alerts? Or some tests/alerts may depend on several major tests (e.g. router/network availability)?
About Windows 2003: IMHO Windows XP and Windows 2003 has more bugs than Windows 200 SP4
Regards
Alex
1) Version of HostMonitor?
2) What kind of alerts do you use? "Send e-mail", "Generate reports" and some other?
3) Do you need to start all these alerts? Or some tests/alerts may depend on several major tests (e.g. router/network availability)?
About Windows 2003: IMHO Windows XP and Windows 2003 has more bugs than Windows 200 SP4
Regards
Alex
Answers to "too many alerts" questions
1. My version is 4.60.
2. The actions giving me problems are "send email". I also generate an hourly report of all failed tests that gets emailed and it is also subject to this problem.
3. Just about all tests/alerts are dependent upon a successful "ping" test.
Do you recommend against Windows 2003 server over Windows 2000 server SP4?
2. The actions giving me problems are "send email". I also generate an hourly report of all failed tests that gets emailed and it is also subject to this problem.
3. Just about all tests/alerts are dependent upon a successful "ping" test.
Do you recommend against Windows 2003 server over Windows 2000 server SP4?
Just tested HostMonitor
- system: Windows 2000 SP4, Pentium 4, 512Mb memory
- created 160 tests
- assigned action profile with "Send e-mail" and "Generate reports" actions. "Generate reports" action creates 3 reports.
- fail all tests at the same time (using "reverse alert" option)
HostMonitor sent 160 e-mails and generated 480 reports (actually 3 reports, 160 times each of report) without any delay and without any problems.
Ok, there is some delay - HostMonitor used 99% of CPU for 5 sec. But everyting works fine. I tested it over and over
May be you are running some programs that hold too much system resources? And system cannot allocate all resources needed by HostMonitor?
Regards
Alex
- system: Windows 2000 SP4, Pentium 4, 512Mb memory
- created 160 tests
- assigned action profile with "Send e-mail" and "Generate reports" actions. "Generate reports" action creates 3 reports.
- fail all tests at the same time (using "reverse alert" option)
HostMonitor sent 160 e-mails and generated 480 reports (actually 3 reports, 160 times each of report) without any delay and without any problems.
Ok, there is some delay - HostMonitor used 99% of CPU for 5 sec. But everyting works fine. I tested it over and over
May be you are running some programs that hold too much system resources? And system cannot allocate all resources needed by HostMonitor?
Yes, I would recommend Windows 2000 SP4Do you recommend against Windows 2003 server over Windows 2000 server SP4?
Regards
Alex
Just a suggestion....
In our setup we have nearly 5,000 tests but alerts are somewhat nested. For example, I may have 20-40 tests on a single machine but none of them will run if the master test (a simple ping) for that machine fails.
So if that machine fails, I get 1 alert (the failed master ping) instead of 40.
Next, these machines are grouped into network segments and each of the individual machine master tests (the pings) are further dependent upon a network segment master test (for example, a ping of the router's interface for that segment). If the router or network switch goes down, I get 1 alert instead of 20 (20 machines in the segment) or worse 800 (20 machines x 40 tests/machine).
It's taken some work, but basically, for whatever problem occurs, I usually get no more than 10 alerts for most problems.
We do a lot of paging when something goes down and there's nothing more annoying than having your pager receive 450 pages....
In our setup we have nearly 5,000 tests but alerts are somewhat nested. For example, I may have 20-40 tests on a single machine but none of them will run if the master test (a simple ping) for that machine fails.
So if that machine fails, I get 1 alert (the failed master ping) instead of 40.
Next, these machines are grouped into network segments and each of the individual machine master tests (the pings) are further dependent upon a network segment master test (for example, a ping of the router's interface for that segment). If the router or network switch goes down, I get 1 alert instead of 20 (20 machines in the segment) or worse 800 (20 machines x 40 tests/machine).
It's taken some work, but basically, for whatever problem occurs, I usually get no more than 10 alerts for most problems.
We do a lot of paging when something goes down and there's nothing more annoying than having your pager receive 450 pages....