Details
L3 - Default Description
Given:
- bootstrap > 4 engines
- each engine connects to the same database
- each engine logs metrics (with different reporter id)
- there exists metrics for more than 24h
When:
Opening the dashboard in Cockpit to see the metrics for the last day
Observed Behavior:
Metrics show 0 at the start of the interval.
Expected Behavior:
Metrics show a consistent workload.
Problem:
The backend has a limitation, as it can only return 200 data points [1]. When Cockpit fetches metrics for the last day, then the resulting array contains 200 data points, but approximately only for the half of the day. The reason is that the data points are grouped by the reporters, so that the data points for the whole does not fit into the array of 200 data points.
Possible Solution:
MetricsQuery has an option to aggregate the values over the existing reporters, so that the resulting metrics are not grouped by the reporters.
