Processor load graphs rounding values

Some graphs are using rounded values even with one hour scale from the UI. For example, a particular CPU-unit was polled to have over 90 % load for a while but the graphs never show this. The value is recorded correctly in the database. It seems that only long lasting high loads are graphed correctly. Is this rounding a configurable option? I would like to use the best possible accuracy (= data from every poll // 5 min) without any rounding when ever possible and when the timescale used is larger, then the graph could use some averages.

For example, this device has the following alert rule: processors.processor_usage > 80 AND macros.device_up = 1. The alert always matches correctly when this threshold is reached but this never shows in the graphs.

During the last poll, this data was recorded in the DB:
| processor_descr | processor_usage |
| FPC: LC2103 @ 0// | 94 |
| FPC: LC2103 @ 1// | 28 |
| Routing Engine 0 | 1 |
3 rows in set (0.001 sec)

But the graph shows a maximum of ~80 % load even with an 1 hour interval.


This topic was automatically closed 730 days after the last reply. New replies are no longer allowed.