graphs for cpu and other not displayed when min step set to auto

[pschonmann]

Hi, im scraping data from servers every 60s.
When i import dashboard in default, every metric where is [$__rate_interval] isnt displayed.
When i set min step to 1m, it works.

Would be possible to add min step as variable ?

ii grafana 11.1.2 amd64 Grafana
hi prometheus 2.24.1+ds-1+b7 amd64 Monitoring system and time series database

[CherrySuryp]

Same here.
Dashboards works just fine when time range set to "last 24 hours".
But when I'm trying to set smaller time range - "no data" error appears in every visualisation where [$__rate_interval] applied.
I fixed it by replacing [$__rate_interval] with [$__range]. But should it work like this?

24h time range applied:

12h time range applied:

[waldner]

Same here, the workaround is also inconvenient because it requires editing a lot of panels.

[Confetti-Camouflage]

Make sure your datasource is configured correctly. Grafana expects the default scraping interval for Prometheus datasources to be 15s. Prometheus, however, defaults to only scraping every 1m. Changing them to both match (in my case Prometheus to 15s) fixed the graphs not showing for me.

[CherrySuryp]

@Confetti-Camouflage Nope, didn't work.
I have scrape_interval: 1m set in prom config, and 1m refresh interval in Grafana. Everything's remains same :(

15s scrape interval seems to be too fast (I have about 200 servers so yeah), so it's easier to update __rate_interval to __range though.

[shaoxp]

i met a similar problem. when use $__rate_interval, lots of panel will meet error as following.

I changed to use __range temporary to go, and it seems everything go fine. not sure whether it has potential risk for this change. please advice if you see anything. thanks.

[pschonmann]

i met a similar problem. when use $__rate_interval, lots of panel will meet error as following.
I changed to use __range temporary to go, and it seems everything go fine. not sure whether it has potential risk for this change. please advice if you see anything. thanks.

Isnt problem, that you replaced "1m" with wrong syntax "1m1" somewhere ?

[CherrySuryp]

Hi everyone, I just found exactly same dashboard template that just works, you might consider switching to it.
https://grafana.com/grafana/dashboards/12486-node-exporter-full/

[igolman]

same here.. I've updated the timeInterval in Grafana Datasources according to the README to the same scrape interval (30s) of prometheus defined for node-exporter job and it worked... Hope this helps..

[rfmoz]

I haven't found any problem with Grafana v11.6.0 and Prometheus 2.53.3.

Anyway, the information available on the README, as @igolman pointed before, could avoid the issue.

[pschonmann]

Same here. Dashboards works just fine when time range set to "last 24 hours". But when I'm trying to set smaller time range - "no data" error appears in every visualisation where [$__rate_interval] applied. I fixed it by replacing [$__rate_interval] with [$__range]. But should it work like this?

Sorry man this leads to bullshit graphs. The solution is set Prometheus source to 1m in my case.