The Metric type picks which number of each stored point the DPS works with. Everything after it (filtering, reduce, resample) only sees that number.
| Metric type | Number used |
|---|---|
| VALUE | val: its meaning depends on the semantics of the metric |
| MAX | highest value recorded in the interval |
| MIN | lowest value recorded in the interval |
| COUNT | number of values recorded in the interval |
The Introscope agent stores one point per 15 s interval, computed from the values recorded in it (an average-style metric here):
| Interval 1 (10:00:00) | Interval 2 (10:00:15) | Interval 3 (10:00:30) | |
|---|---|---|---|
| Values recorded by the agent | 1, 3, 4, 5, 1 | 2, 2, 6 | 10 |
Stored val |
2.8 | 3.33 | 10 |
| Stored max | 5 | 6 |
10 |
| Stored min | 1 | 2 | 10 |
| Stored count | 5 | 3 | 1 |
Data aggregation
The Enterprise Manager keeps the 15 s points for about a week, then merges them into aggregated points of 1 min, then of 15 min, as the semantics of the metric adds up over time:
| Age of the data (default Enterprise Manager configuration) | One point per |
|---|---|
| about a week | 15 s |
| then | 1 min |
| then | 15 min |
- the counts add up, min and max keep the lowest and the highest
Example, an average metric, four 15 s points merged into one 1 min point:
| 10:00:15 | 10:00:30 | 10:00:45 | 10:01:00 | 1 min point | |
|---|---|---|---|---|---|
val |
10 | 20 | no data | 30 | (10×6 + 20×6 + 30×6) / 18 = 20 |
| count | 6 | 6 | 0 | 6 | 6 + 6 + 6 = 18 |
- the count keeps a 1 min point worth three 15 s ones here, when a DPS merges points of both ages
Effective metric only removes the points with count 0 (agent silent) before the metric type is applied.
Time period selector
The Time period selector picks the evaluated period:
-
Use current: the period currently selected in the portal - a saved period: e.g. “Last hour”, or a fixed incident window