Sometimes there are discontinuous indicators

[liuhaibin0528]

I have found that sometimes when certain indicators are converted through YACE, they appear discontinuous on Prometheus, but after a period of time, they return to normal. I don't know what the reason is

etc,AWS/EC2 CPUUtilization metrics
yace config:

prometheus view:

I found that not all instances of this indicator are like this. Occasionally, a certain instance may experience this situation, but there is no pattern

Please give me some guidance, thank you

[liuhaibin0528]

By the way, it's normal on CloudWatch