search

rfmoz / grafana-dashboards

Grafana dashboards
Apache License 2.0
1.17k stars 445 forks source link

parse error: bad duration syntax for CPU and network #116

Closed Elikill58 closed 5 months ago

Elikill58 commented 1 year ago

For multiple charts, I get the error bad_data: 1:101: parse error: bad duration syntax: "5m1" :

image

teamsmiley commented 1 year ago

same here

JannikSommer commented 1 year ago

It seems that changing the $__rate_interval to just $__interval fixes this issue. However, it may be that the stats no longer reflect the data it is designed to.

Elikill58 commented 1 year ago

@JannikSommer Yes, I made a PR about this to suggest the change

firebird631 commented 1 year ago

same here, I've replaced by $range seems $interval does not the job

Nthalk commented 1 year ago

I've found that this is present with prometheus 2.15.x, and is resolved in at least 2.31, I believe I remember seeing a bug report about this being fixed in 2.25.

The most likely issue here is that there's an old prometheus that's running (like my situation).

My particular encounter with this has to do with Ubuntu 20 only having up to 2.15 in the repo, and I have a fixed server running on Ubuntu 22.

There doesn't appear to be a ppa with a more up-to-date prometheus (https://packages.ubuntu.com/search?keywords=prometheus), so a dist upgrade is required or a manual installation.

ZhenyaPav commented 2 months ago

I'm having this issue right now, with the latest prometheus available on Ubuntu 22. A lot of panels, mainly related to CPU, network and disk have errors along the lines of Status: 400. Message: bad_data: parse error at char 125: bad duration syntax: "2m1"