Cisco Cisco Customer Voice Portal Downloads 发行公告
C
HAPTER
3:
A
DMINISTRATION
U
SER
G
UIDE FOR
C
ISCO
U
NIFIED
CVP
VXML
S
ERVER
AND
C
ISCO
U
NIFIED
C
ALL
S
TUDIO
82
a temporary spike. If the average were 90, then this would indicate that the call volume is
very steady.
very steady.
•
maxReqRespTimeInLast
– Returns the maximum time, in milliseconds, it took VXML Server
to produce an HTTP response in the last X minutes where X is entered by the administrator
and when the maximum was reached. A voice browser makes an HTTP request to VXML
Server, which then must respond with a VoiceXML page. Clearly a large response time
would be cause for concern as a slow performing system will cause callers to think that the
application has encountered errors and in extreme cases could cause the voice browser to
time out a request and end a call with an error.
and when the maximum was reached. A voice browser makes an HTTP request to VXML
Server, which then must respond with a VoiceXML page. Clearly a large response time
would be cause for concern as a slow performing system will cause callers to think that the
application has encountered errors and in extreme cases could cause the voice browser to
time out a request and end a call with an error.
•
avgReqRespTimeInList
– Returns the average time, in milliseconds, it took to produce an
HTTP response in the last X minutes where X is entered by the administrator. This value
gives the administrator a good idea of how long it takes VXML Server to handle responses
given the call volume. This could help the administrator decide if the system is overloaded
and beginning to affect the perception of callers regarding the responsiveness of the
application. It also establishes a baseline to compare with the maximum response time. A
maximum response time significantly higher than the average could be an indication that
there is a problem with an external resource accessed by a custom element such as a database
or web service and the few calls that visited that element suffered from bad performance. It
could also help determine if the maximum response time was an isolated event or an
indication of a trend. For example if the maximum response time were 500ms which
occurred near the present, the average was 400ms, the fact that the peak was 500ms is not
alarming, because the average is so high. In this situation the administrator may choose to
throttle down the calls being handled by the system to bring the response times back down to
more acceptable levels.
gives the administrator a good idea of how long it takes VXML Server to handle responses
given the call volume. This could help the administrator decide if the system is overloaded
and beginning to affect the perception of callers regarding the responsiveness of the
application. It also establishes a baseline to compare with the maximum response time. A
maximum response time significantly higher than the average could be an indication that
there is a problem with an external resource accessed by a custom element such as a database
or web service and the few calls that visited that element suffered from bad performance. It
could also help determine if the maximum response time was an isolated event or an
indication of a trend. For example if the maximum response time were 500ms which
occurred near the present, the average was 400ms, the fact that the peak was 500ms is not
alarming, because the average is so high. In this situation the administrator may choose to
throttle down the calls being handled by the system to bring the response times back down to
more acceptable levels.
•
timeoutCallsInLast
– Returns the total number of calls that ended with a timeout in the last X
minutes where X is entered by the administrator. More specifically, this counts calls where
the “result” action of the “end” category is timeout. See Chapter 5: VXML Server Logging in
the section entitled The Application Activity Logger for more on the different results and
how ended values. Under normal circumstances a call should never time out. Many different
types of conditions can yield session timeouts on VXML Server and so knowing if there are
timeouts in the last period of time would tell the administrator how widespread these issues
are.
the “result” action of the “end” category is timeout. See Chapter 5: VXML Server Logging in
the section entitled The Application Activity Logger for more on the different results and
how ended values. Under normal circumstances a call should never time out. Many different
types of conditions can yield session timeouts on VXML Server and so knowing if there are
timeouts in the last period of time would tell the administrator how widespread these issues
are.
•
failedCallsInLast
– Returns the total number of calls that ended with an error in the last X
minutes where X is entered by the administrator. More specifically, this counts calls where
the “result” action of the “end” category is error. See Chapter 5: VXML Server Logging in
the section entitled The Application Activity Logger for more on the different results and
how ended values. This helps the administrator determine how widespread a bug or other
issue that caused a call to end in an error is. For example, if the last 60 minutes yielded only
one failed call, while the issue should be investigated, it may not be a symptom of a larger
more prevalent issue.
the “result” action of the “end” category is error. See Chapter 5: VXML Server Logging in
the section entitled The Application Activity Logger for more on the different results and
how ended values. This helps the administrator determine how widespread a bug or other
issue that caused a call to end in an error is. For example, if the last 60 minutes yielded only
one failed call, while the issue should be investigated, it may not be a symptom of a larger
more prevalent issue.