Event studio as datacenter alarm system

[nimda]

Hello everyone! I'm not really sure which section I should post this question but here goes... I just starting working in datacenter, where Tivoli monitoring and Tivoli common reporting is used.  I have been working with these two products lately and have found Tivoli monitoring situations and workflows too limited for our needs. In Event Studio I can create more complex rules and tasks. Can I use event studio for actively monitoring datacenter environment? Monitoring agents can send data to data warehouse each minute so alarm delay isn't a problem.  What we are trying to accomplish is active alarms and reports combined. For example: if a server has disk I/O higher than threshold, it's detected by Event studio and our storage administrator would receive notification and report of servers disk I/O history, processes with most I/O before actual event and so on...

Is this something event studio can be used for?
What about performance? I  know this can vary a lot based on data and events. Our datawarehouse is about 2TB and growing. Could checking for hundred different events each minute cause performance problems in Event studio/Database servers? Common reporting server has 8CPU and 16GB ram,  Database server(DB2) is at the moment 16CPU and 32GB ram,both has barely any load. 

I'm lost so all help is really appreciated! Thanks ahead!

[CognosPaul]

I'm not sure I like the idea of using Event Studio for this. As you said, you would have to set it up to check every minute. There may be solutions though. Can your monitoring software run applications? You could have it trigger a series of reports, or even an event studio agent.

[nimda]

Thanks Paul. I totally missed triggers in schedule options. Now I can have the best of both products; monitoring software running checks and triggering reports with simple curl command.

[nimda]

I have some problems when implementing this. I need to know which system to run report on, but sending parameters with trigger seems impossible. What I'm trying to achieve is following:

1. Our monitoring system receives an disk usage alert from machine HOSTNAME01 and runs trigger in Cognos server.
2. Cognos runs "Disk usage" report as triggered task with HOSTNAME01 as parameter. 

Problem is that I have no way of sending information about hostname with the trigger. Any other approaches I should look into?

[CognosPaul]

It's been a while since I've done any triggering based scheduling; if I recall, the lack of support for parameters drove me crazy.

Do you have a finite number of servers? You could possibly set up a report view for each system, and schedule each of those with with the hostname as part of the trigger name.

If that's not feasible, there is another solution. Can you insert records into a table? One solution I had was to trigger the report and burst it based on values in a parameter table. The parameter table would feed a pageset, which would populate all of the parameters in the report. Part of the page included a sp to update a flag so as to not call that row again.

Essentially monitoring software inserts n rows into a table containing the hostname and any other details, and triggers the report.
The report is set up on a pageset on that table with flag is null.
An SP in the page of the report goes back to the table and sets a flag to 'run' or whatever.

It's not pretty, but it worked.