Hi, I have telegraf running in 60 sensors with more than 300 inputs running at the same time.
My deployment is with balena and raspberry 4.
I have problems with spikes when the raspberry is using the internet_speed input. For this reason I am trying to schedule my inputs ping. The code looks like this:
[[inputs.ping]]
urls = ["192.168.0.1"]
interval = "3m"
count = 5
ping_interval = 0.25
collection_offset = "1m"
[inputs.ping.tags]
name = "Puerta de Enlace"
70 pings have a collection offset of 1m
70 pings have a collection offset of 2m
70 pings have a collection offset of 3m.
When I do not have this collection offset everything works fine, I have no packet loss and I only have the issue when internet_speed is running (latency spike).
But when I use collection_offset I have this behavior:
I am only showing the latency now, because of the packet loss. you can see that the inputs are working, but the packet loss is a lot as can be seen here:
that Packet Loss is not present when I am doing the same tests without collection jitter.
My scheduling looks like this (yellow rows show the minute they will run every hour):
and my config look like this:
where my metric_buffer_limit = 2000 is defined in balena.
Can anyone tell me if I am doing something wrong please.
