12.2.5 Considerations for defining a status change condition for a monitoring group
By defining a status change condition for a monitoring group, you can monitor the system more precisely from a monitoring tree.
For example, in a system such as described below, where processing loads are distributed using a load balancer, an error on a lower-level node does not necessarily result in a problem in the higher-level monitoring group. In this type of system with special conditions, you can manage the system status more accurately by defining a status change condition for the monitoring group.
Note that the following restrictions apply when you define a status change condition for a monitoring group.
- Organization of this subsection
(1) Examples of defining a status change condition for a monitoring group
In the following example, the load-balancing system shown below is being monitored in a monitoring tree. Load-balancing system in this context means a system that uses a load balancer to distribute processing loads.
|
- The terms used in the explanation below have the following meaning:
-
-
Web system: Monitoring group (Web system)
-
Load-balancing system: Monitoring group (load-balancing system)
-
Server X: Monitoring object (server X)
-
- The following conditions apply:
-
A Web system problem is assumed when the processing loads of 60% or more (three or more) of the five servers that make up the Web system have reached a Warning threshold.
Consider the following approach to relaying the node status when this system is monitored from a tree view.
|
The status change condition in this example is defined as follows:
Node name |
Status change condition for monitoring group |
||
---|---|---|---|
Status |
Child node status# |
Comparison condition |
|
Load-balancing system |
Error |
Warning |
Percentage: 60% or more or Count: 3 or more |
#: The status setting here includes statuses of higher priority. For example, an Error setting includes Emergency, Alert, and Critical statuses.
With these settings, as long as less than 60% of the servers (three of the five servers) are in Warning status or worse, the status of the load-balancing system and Web system remains unchanged from Initial status. Hence, it is not possible to search for status change events from the higher-level load-balancing system or Web system.
If you want to manage status changes in the lower-level monitoring nodes, or to search for status change events in lower-level monitoring nodes from a higher-level monitoring group, we recommend that you define the condition as shown in the following table, for example.
Node name |
Status change condition for monitoring group |
||
---|---|---|---|
Status |
Child node status# |
Comparison condition |
|
Load-balancing system |
Error |
Warning |
Percentage: 60% or more or Count: 3 or more |
Warning or Normal |
Warning |
Percentage: 20% or more or Count: 1 or more |
#: The status setting here includes statuses of higher priority. For example, an Error setting includes Emergency, Alert, and Critical statuses.
(2) Limitations on defining a status change condition for a monitoring group
Bear in mind the following limitations when you define a status change condition for a monitoring group:
-
The status of a child node color-coded as being in Error status will not necessarily be passed to the higher-level monitoring group.
The higher-level monitoring group might remain in Initial status, even if a child node is in Error status. Because the child node status is not passed to the top-level monitoring group, the alarm lamp does not flash.
-
When you search for status change events, some that affect lower-level nodes might be missing from the results.
If there is a monitoring group in Initial status between the monitoring group from which you are searching and the monitoring object that has Error color-coding, status change events below the group in Initial status will not be retrieved.
In this case, perform a monitoring node search to find nodes that have Error color-coding, and then perform a search for status change events.
If these two limitations are likely to be issues, define the condition so that any one child node in Error status will change the higher-level monitoring group to Warning status, for example.
-
You must review the status change condition if child nodes are added or deleted and their number changes.
You must review the definition if the number of child nodes increases or decreases. For example, if a status change condition is set for a monitoring group of five child nodes, the count (3 or more) and the percentage (60% or more) mean the same. However, they mean different things when the number of child nodes increases, as follows:
-
Count: 3 or more. If another five child nodes are added, making a total of 10, the count will be unchanged. (The status of the group changes when three nodes are in the specified status.)
-
Percentage: 60% or more. If another five child nodes are added, making a total of 10, the percentage will be 60% of 10; that is, six. (The status of the group changes when six nodes are in the specified status.)
The Central Scope does not automatically redefine status change conditions. You should therefore periodically review the condition definitions.
-
-
When the completed-action linkage function is enabled, the status of the monitoring group changes to Initial when you finish taking action on all the lower-level monitoring objects (when you change all status change events to Processed status). That is, the monitoring group will not be searched when you perform a status-change event search.
If this limitation is likely to be an issue, define the condition so that a status change in any one child node cause a status change in the monitoring group. For example, specify that when a child node shifts to Normal status, the monitoring group changes to Normal status.