How status is computed
This topic explains how the status of an application is computed. The status takes into consideration the severity of each node and thus determines the severity of the entire application. There are two types of algorithms used to compute the status: highest and quorum.
The following is a list of severity types in descending order of severity:
Status | Icon | Description |
|---|---|---|
Critical (red) |
| At least one monitored metric exceeded the defined Critical threshold. |
Major (dark orange) |
| At least one monitored metric exceeds the defined Major threshold. |
Minor (light orange) |
| At least one monitored metric exceeded the defined Minor threshold. |
Warning (light orange) |
| At least one monitored metric exceeded the defined Warning threshold. |
Information (blue) |
| Information about the item |
OK (green) |
| No monitored metrics exceed the defined thresholds. |
Every node in an application model has a status by default. If no events are associated with a node, then the status of the node is OK. If there are events associated with a node, either the highest severity algorithm or the quorum algorithm is used to calculate the status of the node. The highest severity algorithm is used by default.
Highest severity algorithm
By default, the status of every node is calculated using the highest severity algorithm. This algorithm finds the highest severity among all the events associated with a node and assigns that severity as the status of the node.
Algorithm/Formula
Status = Highest severity of impacting events on the node
Quorum Algorithm
The quorum algorithm is applied to calculate the status of a node only when a node is of type cluster or group. Only when a node is marked as group or cluster, can a quorum value be specified for the node. The quorum value is always specified as a percentage value. The quorum algorithm finds the severity of all the nodes in the cluster and uses the following formula to compute the quorum. The status is then computed as the lowest severity among the nodes according to the number derived from the formula.
Algorithm/Formula
Status = Lowest severity among the number of cluster elements with highest statuses
Number of cluster elements with highest statuses = quorum_percent * (Number of hosts /100)
Example
Consider a cluster node comprising 4 hosts: host 1, host 2, host 3, host 4. The severities of the hosts are as follows:
host 1=OK, host 2=OK, host 3=MAJOR, host 4=MINOR
Assume that the quorum percentage is 50.
As per the quorum algorithm, number of cluster elements with highest statuses = quorum_percent * (Number of hosts /100)
That is, number of cluster elements with highest statuses = 50 * (4 /100) = 2
Status = Lowest severity among the number of cluster elements with highest statuses
That is, status = lowest severity among the 2 cluster elements with highest statuses = [host 3=MAJOR, host 4=MINOR]
Status = MINOR
Sample example to compute severity
The following table lists different scenarios in an application and determines how the status is computed in these scenarios.
Node | Type | Cluster | Status propagation | Quorum | Contents | Dependent nodes | Events | Status | Propagated status = | Overall status = |
|---|---|---|---|---|---|---|---|---|---|---|
Application 1 | Application | No | N/A | N/A | N/A | Server 1, Server 2 | NIL | OK | Critical | CRITICAL |
Server 1 | Device | No | ON | N/A | host1 | Web Cluster, Server 3 | NIL | OK - No events associated with it | CRITICAL | CRITICAL |
Server 2 | Device | No | ON | N/A | host2 | DB Group, Server 4 | 1 INFO event | INFO- Max (event severity) | MAJOR | MAJOR |
Server 3 | Device | No | ON | NA | host3 | NIL | 1 CRITICAL, 2 MAJOR | CRITICAL - Max (event severity) | NA - No dependent nodes | CRITICAL |
Server 4 | Device | No | ON | NA | host4 | NIL | 2 MAJOR, 2 MINOR | MAJOR - Max (event severity) | NA - No dependent nodes | MAJOR |
Web Cluster | Group | Yes | ON | 50% | host5, host6, host7, host8 | NIL | host5 - 2 MAJOR | INFO | NA - No dependent nodes | INFO |
DB Group | Group | No | ON | NA | host9, host10 | NIL | host 9 - 2 INFO, 1 WARNING | WARNING - Max (event severity) | NA - No dependent nodes | WARNING |
Related topic