Value: [no value] In Grafana Resolve Alert Notification

I have Grafana charts showing NetApp Storage Disk Utilization.
I have setup alert, if Disk utilization is 99% for 5 mins, send alerts via Slack Channel.

When the disk utilization related to a RAID Group meets this condition, an alert pops in Slack channel with the name of the RAID group having high disk utilization. However, when the condition gets resolved, the Alert in Slack channel shows Value: [no value].

[FIRING:1] Disk Utiliziation ASE | NetApp Harvest 21.11 - cDOT (critical ontap)
Firing
Value: [ var=‘B0’ metric=‘node01 - n1_aggr1’ labels={aggr=n1_aggr1, node=node01} value=99.689819061119 ]
Labels:

  • alertname = Disk Utiliziation
  • grafana_folder = ASE | NetApp Harvest 21.11 - cDOT
  • severity = critical
  • storage = ontap

[RESOLVED] Disk Utiliziation ASE | NetApp Harvest 21.11 - cDOT (critical ontap)
Resolved
Value: [no value]
Labels:

  • alertname = Disk Utiliziation
  • grafana_folder = ASE | NetApp Harvest 21.11 - cDOT
  • severity = critical
  • storage = ontap

There are multiple RAID groups on this storage system. With the RESOLVED message i can’t identify which RAID Group has this condition resolved.

Any help is much appreciated

Welcome @nitishchopra to the Grafana forum.

In the text below, what is the name of the RAID group?

Thanks @grant2
Raid Group/Aggregate name: n1_aggr1

What is your datasource?

If Prometheus, can you share your screen that looks like this?

If something else, can you share your A, B and C sections (like this)?

@grant2

Thanks for your post.
We use Prometheus as the datasource.

Hi!

Before I explain what’s happening here, it sounds to be like you want to be using multi-dimensional alerting via either a Reduce + Math or Threshold expressions, rather than Classic Conditions.

In Classic Conditions, each alert rule can fire at most one alert. If there is more than one series that has exceeded the condition, then these are shown in the ValueString as can be seen in the first example.

If there are two series that had exceeded the condition, but now one of them has resolved, you won’t get a resolved notification because there is still one series that has exceeded the condition. The alert will still be firing, just the text in the Value field will change at the next repeat_interval.

If all series have resolved then the alert will also be resolved. [no value] means that there are no firing metrics for the Classic Condition.

It sounds to be that you would like to get a notification which each series exceeds the condition and is resolved. This is where I would recommend using a Reduce + Math expression, or a Threshold expression, instead of a Classic Condition, and in doing so allow you to use multi-dimensional alerting.

1 Like

For Promoetheus Multi-Dimensional alerting, there is a great video here.

1 Like

@georgerobinson and @grant2
Thanks for your help.

I have updated the alert with multi-dimensional alerting and got the designed results

The Alert rule is setup as below

The looks good to be rolled out to production.

Thanks

Hello.
Thanks, I have followed the suggestions (reduce + math expression) but the problem persists, below are the screenshots of the rule and the emails with the firing and resolved status. C is the alert condition.
The version of grafana that I am using is: 9.5.7
and the datasource is InfluxDB 2.6.1

RULE:

mail: firing

mail: resolved