[🐛 Bug]: Correlation Incident prefix intermittently ignored

[hrabalvojta]

Describe the bug
Correlation Incident prefix is not included in all Incident names.

So I ended up with this:

Status	Incident	Alerts	Sources
Firing	INC-9 - KubeDeploymentRolloutStuck	1	prometheus
Firing	KubePodNotReady	2	prometheus
Firing	INC-7 - KubeDeploymentReplicasMismatch	1	prometheus

To Reproduce
Steps to reproduce the behavior:

0. Have Kubernetes cluster with kube-prometheus-stack deployed
1. Configure provider prometheus
2. Configure alertmanager to send alerts to keep
3. Create simple manual correlation

Correlation name: default
Append to the same Incident if delay between alerts is below: 72 hours
Incident name template: {{ alert.alertname }}
Incident prefix: INC
Select attribute to group by: alert.alertname
Resolve on: all alerts resolved
Start incident on: Any condition met
Alerts threshold: 1
Filter:
  Source: prometheus

4. Create fail deployment to create alerts
   I created deployment with non-existenting image spec.template.spec.containers[0].image: afdahf and spec.replicas: 2
5. When alerts are triggered you should see incidents above.

Expected behavior
It is expected that correlation - Incident prefix will be present in all incidents.

Status	Incident	Alerts	Sources
Firing	INC-9 - KubeDeploymentRolloutStuck	1	prometheus
Firing	INC-8 - KubePodNotReady	2	prometheus
Firing	INC-7 - KubeDeploymentReplicasMismatch	1	prometheus

Screenshots
If applicable, add screenshots to help explain your problem.

Additional context
Versions:

• kubernetes: v1.34.1+k3s1
• keephq helm chart: keep-0.1.93
• kube-prometheus-stack helm chart: kube-prometheus-stack-78.5.0