Grafana managed alerts are so fucking over complicated it surprises me that it is a professional product.

And if you need help, the grafana community forum is a fucking ghost town

And the docs suck too

what's so complicated about it?

Don't rely on grafana to do that.

Grafana is in essence just a visualisation tool. It just show fancy shmancy graphics on metrics. That's all!

Prometheus is better in it's alerting system.

@netikras here is an example. I want a notification when a substring appears in a log.

Grafana managed alerts requires that the the condition to trigger the alert happens TWICE within a specific time frame (which can’t be less than 10s) for the notification to be sent.

So if that substring only appeared in a single log, then your not getting a notification

FUCK

@DeepHotel wat? Why twice? Worked with just one of them for me

@netikras i have no idea lol. Could be my own incompetence. From what I understand, when the alert rule is true, the alert will be “pending”. If the alert is still true while pending, the notification will be sent

@DeepHotel

https://grafana.com/docs/grafana/....

tried it just now in Grafana Cloud

@netikras huh, I’m using grafana OSS so maybe it’s different. What’s your evaluation interval? And for each rule in that evaluation group, with a the # Evals?

I thought the pending interval couldn’t be smaller than the rule group’s evaluation interval but it seems like I’m wrong with the docs u sent me