contradicting Grafana curves

[jochen-schuettler]

What happened?

The consumption curve in kWh shows several large steps between 12 and 1 p.m. The consumption curve in W shows peaks only at one time roughly 12:30.

What did you expect to happen?

Either there is one kWh step at the same time as the peak in W, or there must be several peaks in W at the times when there are steps in kWh.

How can we reproduce it (as minimally and precisely as possible)?

I have no idea.

Anything else we need to know?

Kepler image tag

0.6.1

Kubernetes version

1.24

Cloud provider or bare metal

AWS EKS / ECS

OS version

Amazon Linux 2

Install tools

Helm

Kepler deployment config

unchanged from given Helm chart

Related plugins (CNI, CSI, ...) and versions (if applicable)

CNI Version: v1.12.0-eksbuild.2 CSI Version: v1.14.0-eksbuild.1

[rootfs]

not sure about the first issue, but for the 2nd (i.e. decreasing curve), since it is on a 24 hour period basis (i.e. not accumulative) the drop makes sense.

[jochen-schuettler]

I deleted the 2nd non-issue. Do you have any ideas how to gain more info on the repeating problem?

[jochen-schuettler]

I updated the Details now

[jochen-schuettler]

An idea: If the kWh-curve is smoothed by an averaging filter, that could explain the behaviour. Is that the case?

[marceloamaral]

Yes, KWh extrapolates the average values over a longer interval.