Help to understand ICMP monitoring

[TheFactor82]

Hi everybody,
I'm using the ICMP Ping Template to monitor some hosts of my infrastructure. As far as I know, the "ICMP response time" (that uses key icmppingsec as a Simple check, type FLOAT), gets values in SECONDS.
So, I think that it is quite normal that all my values are 0.
What I don't understand is this:

this is a graph with a time range of 1h. It's what I expect: a line with all 0




this is the same graph with a time range of 7days. The line is around 200ms, and I can see the values collected (198.3ms, 470.6ms ...).

Someone can explain this behaviour?

[TheFactor82]

it's just a problem of cutting screenshot, the second screenshot goes also to date 3.19.
However, the problem is not the comparison. The question is: how can trends calculate an average of 200ms if the collected values are all 0's ???