Create groups and dependencies

You can group multiple individual monitors together to assign dependencies. Properly-configured groups and dependencies act as an alert suppression system in ipMonitor. When a critical resource fails, ipMonitor limits alerts to the monitor defined as a dependency rather than triggering alerts for every member monitor in the group.

Groups consist of two monitor types: Group Members and Group Dependencies. Group Members are monitors that make up the group. Group Dependencies are monitors that must remain available for all members to function properly.

Dependencies define the relationship between critical resources and resources that depend on them for all or part of their functionality. A router, switch, server computer, or stable path to another network would all be valid dependencies.

Defining dependency relationships allows you to:

  • Minimize the number of redundant alerts
  • Isolate the root cause of the problem
  • Prevent configured recovery alerts from attempting to restart services and applications if such an action is not required

For example, consider a web solution comprised of various individual components. It would be practical to group them together and assign certain dependencies to the group.

A logical dependency would be the network switch. If the switch fails or becomes unavailable, all member monitors for both groups would be affected, generating several alerts.

When you define the switch monitor as a dependency, only the switch failure triggers an alert. This helps you:

  • Isolate the root cause of the problem
  • Minimize the number of unnecessary alerts
  • Ensure that recovery alerts are not triggered for services and applications that do not need to be restarted

Dependency monitors must reach the Down state for ipMonitor to disable alerting for group members. If the dependency monitor is Up, individual group members can still trigger alerts. For example, alerts will be processed if hard disk space runs low, an SNMP trap is received, the SQL server produces an unexpected query result, and so on.

After a dependency monitor reaches the Down state, members that are currently in an Up or Warn state will not reach Down or Lost state. When all dependencies return to an Up state from a Down state, all alerting functionality is restored.