The dashboard [OTEL][Metrics Kubernetes] Cluster Overview is returning errors.

### Integration Name: Kubernetes OpenTelemetry Assets [kubernetes_otel]

### Integration Version: 1.4.0

### Agent Version: v9.2.5

### Elasticsearch Version: v9.2.5

### OS Version and Architecture: v9.2.5

**Summary**
OpenTelemetry default dashboards fail to render *Workload Overview* visualizations (Pods, Deployments, DaemonSets, StatefulSets, ReplicaSets) when used with cross-cluster search (CCS).

**Environment**

* Elastic Stack: 9.2.5
* Deployment: Elastic CLoud
* Integration: Kubernetes OpenTelemetry Assets
* Dashboard: [OTEL][Metrics Kubernetes] Cluster Overview

**Issue**
In CCS setups, several visualizations (notably Workload Pie charts) return errors (see attached screenshot). The same dashboards work correctly when used within a single cluster.

This appears to be caused by ad hoc data views generated by the dashboards using the index pattern `metrics-*otel*`. These temporary data views do not resolve correctly in CCS (missing `:metrics-*otel*
**Impact**
Since ≥ 9.2.0, Kubernetes OTLP metrics are not fully available in the Observability Infrastructure UI, increasing reliance on these dashboards. As a result, this issue significantly impacts usability in KIbana environments.

**Expected**
Dashboards and ad hoc data views should also support remote cluster index patterns when dashboards are running in Cross Cluster Search. (for example it should also cover remote_cluster:metrics-*otel*)

**Actual**
Visualizations fail due to incorrect index patterns (`metrics-*otel*`) in temporary data views.
Same issue also exist by logs which is using for Cluster Events visualiosations.

**Attachments**

* Screenshot showing visualization errors in Workload Overview dashboard

### What did you do?

I tried customizing the dashboards, but after updating the integration or the cluster, the changes get overwritten and the incorrect version is imported again.

### What did you see?

In CCS setups, several visualizations (notably Workload Pie charts) return errors (see attached screenshot). The same dashboards work correctly when used within a single cluster.

This appears to be caused by ad hoc data views generated by the dashboards using the index pattern `metrics-*otel*`. These temporary data views do not resolve correctly in CCS (missing `:metrics-*otel*
**Impact**
Since ≥ 9.2.0, Kubernetes OTLP metrics are not fully available in the Observability Infrastructure UI, increasing reliance on these dashboards. As a result, this issue significantly impacts usability in KIbana environments.

### What did you expect to see?

Dashboards and ad hoc data views should also support remote cluster index patterns when dashboards are running in Cross Cluster Search. (for example it should also cover remote_cluster:metrics-*otel*)