EMS Manual

Standard Events
EMS Manual—426909-005
9-31
Usage Threshold Event Tokens
For more information about the use of this event, see Requirements for Standard
Events on page 9-3 and Proactive Problem Management Functions on page 9-8.
These event-specific tokens must be provided in addition to the common-standard
tokens. Unconditional token types are identified with a “U.”
Usage Threshold Event Tokens
A Usage Threshold event is reported for a resource whenever the usage of the
resource crosses a preconfigured threshold level. Usually, the subsystem that controls
the resource (one that can take the resource in or out of service), or the subsystem
that uses the resource must report this event. But it could also be the program that
monitors the resource that must report this event. For more information about the use
of this event, see Proactive Problem Management Functions
on page 9-8.
The utilization level of a system resource is intended to supply a measure of how much
an object is being used. The level could be specified as a percentage (between zero
and one hundred) or as an absolute number. Subsystem developers are free to choose
the best way to represent the utilization data of their objects.
If the level is specified as a percentage, zero means the object is absolutely idle over
some period of time, and 100 means the object was never idle over the same period of
time. The period of time for measurement (sampling time) that is meaningful is
determined by the subsystem developer.
For certain objects, the utilization level is more appropriately specified in an absolute
number rather than a percentage. For example, transaction response times are
measured in units of time, queue lengths in units of requests, storage in units of bytes,
and rate in units of bytes per second.
The subject of this event is the object whose usage threshold is being reported. If
multiple objects experience the same changes and all have the same configured
Type of
Fault
ZEMS-TKN-TXFAULT-TYPE (ZSPI-TYP-ENUM,U)
Enumerated values are defined by subsystems, and they must be greater
than or equal to ZEMS-VAL-MIN-USER-VALUE.
Indicates the type of transient fault like correctable memory error, excessive
line resets, or recoverable controller errors.
Private
Tokens
Additional event-specific tokens should be provided by subsystems
and applications that could help the operator diagnose the transient
fault.
For example, for correctable memory error, the diagnostic information could
be the hardware address of the memory chip that experienced the fault and
the error code if known. For retryable controller error, the diagnostic
information could be the error data that the controller returns to the I/O
process like SNAX/XF or X25AM. For excessive network resets, the
diagnostic information could be the error data provided by the network and
the excessive reset count.