The availability of applications depends on the availability of various application components like server, databases, etc. and the network. So when your application goes down, how do you know if it is a network problem or an application problem?
Simple, just configure the router as a dependent device to the business application and suppress the alarm for the dependent monitor.
This will ensure that you will receive alerts only for the unavailability of the router/network device and not for your business application. This way the network team could be notified of the problem without interrupting the application or system admin team.
You can also associate multiple dependent devices across managed servers and suppress alarms from the dependent monitors. This gives you more flexibility in associating dependent parameters for complex applications and helps you isolate the root cause of performance problems in a better way.
Featured White Paper
Integrated Performance Management for Physical, Virtual and Cloud Infrastructure
Find out the performance management strategies that IT departments should adopt to overcome new application performance management challenges and to ensure their services meet business goals.
Other Options to Reduce False Alerts
1) Consecutive Polls before Reporting an Error
You can specify the number of times Applications Manager should poll consecutively before generating critical, warning or clear alarms. This option can be specified at the global level for all monitors, at the threshold level as well as the individual attribute level. Read more >>
2) Check for Network Availability
If your Applications Manager instance loses network connectivity, you might receive a flood of down alarms for monitors. To avoid alarms of this type, just enable the Check for Network Availability option under the Admin->Availability Settings page. Read more >>
3) Check for URL Availability
If your external proxy settings are wrongly configured in Applications Manager or if Applications Manager gets isolated from the network, all your website monitors will be shown as down. To avoid these type of alarms, enable the URL availability check. This will ensure that Applications Manager will check for the availability of a reference url before sending alarms for downtime. Read more >>
4) Enable Ping Retries
By default, Applications Manager sends out Ping requests to a server only once. If this request does not fetch a response from the server, the server is declared as down. With the 'Ping Retries' option, you can specify the number of times Applications Manager should Ping a server before it is declared unavailable. This option can be specified in the AvailabilityTests.conf file under <AppManager>/conf directory.
5) Enable Port Check
Sometimes Applications Manager will be unable to get a response for its Ping requests as the network may discard ICMP packets. To avoid alerts due to this problem, you can enable the Port Check option in the AvailabilityTests.conf file under <AppManager>/conf directory. Once this is enabled, Applications Manager will connect to the ports in the server and send out alerts only if the ports are not reachable.
From the Community
Here are some useful discussions from our user forums and blogs:
MDB Pool Counts under EJB details for Websphere 6.1
Purge Database (MSSQL)
Enable URL Availability Check to Reduce False Alarms for Website Monitors