HP 7102dl - ProCurve Secure Router Configuration Manual page 443

You can set the tolerance in one of two ways:
consecutive failures—If a certain number of tests in a row fail, the
probe fails.
With this type of tolerance, the probe counts consecutive failures. Any
time a test passes, the probe resets the count. When the count reaches the
tolerance setting, the probe fails.
rate of failure—If tests fail at a certain rate, the probe fails. You specify
the rate by configuring two settings: the number of tests that can fail
within a set of tests, and the number of tests in the set. For example, you
can specify that the probe enters the Fail state if 20 out of 25 tests fail.
The size of the set determines the number of tests that the probe considers
and is a running count. For example, if the set size is 25, the probe counts
all failures in only the 25 most recent tests.
Note that, even after a probe has entered the Fail state, it continues to run
tests. In this way, the probe can revert to the Pass state when the monitored
conditions become acceptable again.
The conditions for a probe passing once again also depend on the type of
tolerance. A probe with a consecutive failure tolerance reverts to the Pass
state as soon as one test passes. For a probe with a rate of failure tolerance,
reverting to the Pass state is slightly more complicated. Enough tests must
pass that, when looking at the X most recent tests (X being the set size), the
number of failures does not exceed the number set by the tolerance.
The consecutive failure option is the most straightforward, particularly when
you simply want to test connectivity. You might choose the rate of failure
option when setting up a probe to monitor the performance of a server. When
a probe sends a series of requests to an over-burdened server, chances are that
at least one request will get through, so a probe with a consecutive failures
tolerance might detect a problem rarely, if ever. However, the server could still
be failing to respond an unacceptable percentage of the time. A rate of failure
tolerance would detect the poor performance.
To set the tolerance according to consecutive failures, from the network
monitor probe configuration mode context, enter:
Syntax: tolerance consecutive-failures <failures>
The valid range for failures is from 1 to 255.
To set the tolerance according to rate of failure, from the network monitor
probe configuration mode context, enter:
Syntax: tolerance rate-of-failure <failures> <set size>
Network Monitoring
Configuring Network Monitoring
9-15