Hi guys!
I need your help with those Grafana-managed alert rules. I just cant get my head around on how I get them to work as I expect it and the Documentation doesnt seem to give me much insight either. Hopefully some of you can tell me what I did wrong or missing out.
What do I want/need?
So I have multiple sensors feeding data into my InfluxDB and I would like to get a message, if one of them suddenly stopped broadcasting and of course when it comes back on.
Here are screenshots of how my rule is configured at the moment:
What do I expect from this alert rule?
I expect it to check once every hour, if in the last 2min the “frequency” value was outside of range 45-55. If yes, fire an alert+send mail, if no, do nothing. When value comes back in range, resolve the alert+send mail. If there is no data, fire an alert as well, if data comes back in, resolve alert and send a mail for both cases.
I realised that the “Pending period” is a crucial setting, but even when I set it to “0” the alerts do not come as expected, sometimes they fire or resolve hours later. Sometimes right away.
If the health state shows “nodata” the alert state stays at “Normal” which shouldnt be for the sake of my configuration and normal human thinking. Anybody knows why?
So what am I doing wrong or missing out?
Do I have to use Loki or Prometheus to reach my goal?
Is it possible to adjust the rule so it just looks out for data points and no specific values?
I am not an studied IT-guy and working my way from scratch into this, so please be patient with me^^
Thank you to everybody taking the time to read and answer :*
Happy 2025 for everybody!