About threshold settings
Using InfoScale Operations Manager 9.1, you can set the threshold values on the respective objects for the metrics that are described in Table: Objects, metrics, and evaluation intervals. A Risk or Fault is raised when the threshold condition is violated.
Managed host version 7.2 or later is required to enable threshold settings on the host and is supported only from one Management Server. Threshold settings are not supported on an agentless host. You can view the configuration type of the host in the Management Server perspective.
Managed host versions 7.2 have the following default threshold values which are not displayed on the Management Server console.
-
CPU utilization = 98 %
-
Available Memory = 102400 KB
An alert of the risk severity type is raised if these default threshold values are violated. However if you define new threshold values on a managed host version 6.1, then the default values are overwritten.
The following table lists the objects and the metrics on which you can set the threshold value. It also lists the evaluation intervals for the objects.
Related information
Table: Objects, metrics, and evaluation intervals
| Object | Metrics | Evaluation intervals |
|---|---|---|
| Host | CPU Utilization | 5 minutes |
| Available Memory | ||
| Average CPU Load | ||
| Disk, volume, and path associated with a host. | Average Read Latency | 5 minutes |
| Average Write Latency | On a UNIX/Linux host, the performance statistics for disk and volume are collected every one minute. Last five samples are considered for threshold evaluation and if any of these samples violate the threshold value, an alert is raised. | |
| Host Initiator | Average Read Latency | 5 minutes |
| Average Write Latency | Note: Threshold setting for host initiator on a Windows host is not supported. | |
| Read Queue Length | ||
| Write Queue Length | ||
| Read Errors | ||
| Write Errors | ||
| Cluster and service group | Failover Duration | Threshold evaluation happens on failover event. |
Sample values are collected for the performance metrics for an object at the interval mentioned in Table: Objects, metrics, and evaluation intervals. If any of the collected sample values violate any of the threshold values defined for that metric, an alert is raised with severity specified in threshold definition.
The alert is cleared only when the collected performance sample value does not violate any of the threshold values defined for that performance metric.
More Information