Scope PrometheusRule to a specific TempoStack instance

[andreasgerstmayr]

When enabling .spec.observability.metrics.createPrometheusRules multiple alerts are generated, for example:

      - alert: TempoDistributorUnhealthy
        annotations:
          message: There are {{ printf "%f" $value }} unhealthy distributor(s).
          runbook_url: https://github.com/grafana/tempo/tree/main/operations/tempo-mixin/runbook.md#TempoDistributorUnhealthy
        expr: |
          max by (cluster, namespace) (tempo_ring_members{state="Unhealthy", name="distributor", namespace=~".*"}) > 0
        for: 15m
        labels:
          severity: warning

These alerts match Tempo instance of the entire cluster, but they should only match Tempo instances of the TempoStack where createPrometheusRules was enabled.

Suggested solution: add cluster="$tempostack_name", namespace="$namespace" to every alert.
Note: Also update the recording rules.