Helm: default dashboards use a job label that doesn't exist

[alarobric]

Describe the bug The default Loki dashboards shipped with the Helm chart fail to load data, due to using a job label that doesn't match the metrics that are collected. The dashboard is querying metrics like this: loki_ingester_memory_chunks{cluster="$cluster", job=~"$namespace/(loki|enterprise-logs)-write"} but the ServiceMonitor is generating metrics with job=namespace/loki, not `job=namespace/loki-write'

To Reproduce Steps to reproduce the behavior:

1. Installed loki in EKS using the Helm chart. Installed promtail, grafana, and kube-prometheus-stack alongside.
2. Open the 'Loki/Chunks' dashboard for example. No data is shown.
3. Drill into a chart widget, and modify the query to remove -write from the job label. Data shows as expected.

Expected behavior Default dashboards should work alongside the default ServiceMonitor configuration.

Environment:

• Infrastructure: EKS
• Deployment tool: Helm via Terraform

Screenshots, Promtail config, or terminal output

[alarobric]

I guess this sounds like a related issue to #7657 or #9469