Manage and Configure Events

Contents
casp1032
Events
To illustrate
DX NetOps Spectrum
advanced event processing, consider the following scenario. For testing purposes, you want to temporarily allow a high child count. You want to generate an alarm only if the high child count condition persists for more than one minute, and to clear any such alarms when the watch is reset. Using an EventPair Event Rule,
DX NetOps Spectrum
can perform the following tasks:
  • Generate an initial event when the ChildLimit watch is violated.
  • Suppress the alarm unless that initial event is
    not
    followed by a reset event within 60 seconds.
  • Generate another event when all instances of this watch are either reset or deactivated.
For more information about event rules and creating events, see Event Configuration.
Configure Events
The following procedure is an example of event configuration. Building from the example in the previous topic, the first event is "Initial child count too high" (0xffffffff), and the second event is "Reset" (0xfffffffe). If the child count threshold is violated and is not reset within 60 seconds, a "Child count too high" alarm (event/alarm 0xfffffffd) is generated.
Follow these steps:
  1. Select Utilities, Event Configuration from the Tools menu.
    The Event Configuration page appears with the Navigation, Contents, and Details panels.
  2. Click create_event_icon (Create Event) in the Navigation panel.
    The Create Event dialog appears, with Event Code populated with a default value.
  3. (Optional) Edit the event code and enter an event message and click OK.
    The event is created and its details appear in the Contents and Details panels on the right.
  4. Create and save the initial Child count too high event (0xffffffff).
    This event triggers the EventPair rule that conditionally generates an alarm and a Reset event (0xfffffffe). The watch uses the Reset event to evaluate the persistence of the child count condition and to clear any alarms that the EventPair rule generates when the watch is reset.
  5. Create and save an event "0xfffffffd" and a corresponding alarm that is used to generate a notification if the child count is persistently high.
  6. Create and save an Event Pair rule for the initial "Child count too high" event (0xffffffff) that generates the "0xfffffffd" event and alarm if the Reset event (0xfffffffe) is not received within 60 seconds.
    If the
    SpectroSERVER
    is updated while an event is being processed using an Event Rule, the processing is not completed. All event rules that are processing when this functionality is invoked are flushed.
  7. Click File, Save All.
    The new events are saved.
  8. Open a watch for editing.
  9. On the Threshold tab, select Generate Event(s) under Notification.
    The event code fields appear.
  10. Enter the initial "Child count too high" event (0xffffffff) that you previously created in the 'when threshold is violated' field.
  11. Enter the alarm-clearing "Reset" event (0xfffffffe) that you previously created in the 'when threshold is reset' field, and click OK.
    The event is associated with the watch.
    If the LAN_802_3 container contains 5 or more models, the initial "Child count too high" event is generated, and the event rule is evaluated. An alarm is generated only if the reset event is not received within 60 seconds of the initial event. If the watch is reset, any "Child count too high" alarms that the watch generates are cleared.
The EventDisp entries (if needed) for this example are as follows:
# make sure child count too high (0xffffffff) is # reset (0xfffffffe) within 60 seconds # otherwise generate alarming event (0xfffffffd)0xffffffff E 50 R Aprisma.EventPair, 0xfffffffe, 0xfffffffd, 60 # generate an alarm on persistent child count too high0xfffffffd E 50 A 1,0xfffffffd # if child count drops below threshold, clear any existing alarms0xfffffffe E 50 C 0xfffffffd
Event Codes
Event codes are generated in the
DX NetOps Spectrum
Event Log for each event.
An event is not generated for watch creation.
The following list contains event codes and their descriptions:
  • 0x0048000a
    Watch could not be activated for a model
  • 0x00480000
    Modification of a watch
  • 0x00480001
    Destruction of a watch
  • 0x00480002
    Activation of a watch
  • 0x00480003
    Deactivation of a watch
  • 0x00480004
    Violation of a watch threshold
  • 0x00480005
    Resetting of threshold status for a watch (with variable reset value)
  • 0x00480006
    Violation of an instance of a watch
  • 0x00480007
    Resetting of threshold status for an instance of a watch (with variable reset value)
  • 0x00480010
    Bad probable cause message
  • 0x00480012
    Logging interval conflict
  • 0x00480013
    Error executing a script (for threshold watches only)
  • 0x00480014
    Resetting a watch threshold for a model, with reason
Event Format Files
You can use the existing event format files Event00480004 and Event00480005 (located in <
$SPECROOT
>/SG-Support/CsEvFormat) as templates to create your own watch-related event format files.
For more information, see
Event Variables
The events for watch threshold violation (0x00480004) and reset (0x00480005) have variable bindings. This applies both to watches that generate alarms directly and watches that generate events. The clear event has a single variable binding, ID #2, containing the watch name.
The following table shows the binding variables and their contents:
Variable Binding ID
Contents
2
Watch name
4
Threshold reference value
5
Threshold reset value (only for reset events)
6
Comparison string value
7
Calculated watch value
8
Instance object identifier (OID) if the watch has instances, otherwise empty
9
Internal; strings to complete the respective event message
10
Internal; strings to complete the respective event message
11
Internal; strings to complete the respective event message