rolling-max
Input Stack: |
⇨ |
Output Stack: |
Maximum value within a specified window. This operation can be used in
alerting expressions to find a lower bound for noisy data based on recent
samples. For example:
name,sps,:eq,:sum,
:dup,
5,:rolling-max
Missing values, NaN
, will be ignored when computing the min. If all values
within the window are NaN
, then NaN
will be emitted. For example:
Input |
3,:rolling-max |
0 |
0 |
1 |
1 |
-1 |
1 |
NaN |
1 |
0 |
0 |
1 |
1 |
1 |
1 |
1 |
1 |
1 |
1 |
0 |
1 |
The window size, n
, is the number of datapoints to consider including the current
value. Note that it is based on datapoints not a specific amount of time. As a result the
number of occurrences will be reduced when transitioning to a larger time frame that
causes consolidation.
Since: 1.6