Is your feature request related to a problem? Please describe.
The current Grafana dashboard for File.d has hard-coded datasource UIDs and no templating for job or instance.
As a result, importing the dashboard into another environment requires manual edits, and all jobs/instances are mixed together in queries, producing noisy or confusing metrics.

Describe the solution you'd like
Make the dashboard portable and configurable by:

• adding a Prometheus datasource input (DS_PROMETHEUS) and __inputs section;
• replacing hard-coded datasource UIDs with ${DS_PROMETHEUS} references;
• introducing templating variables for job, instance, and pipeline with proper dependencies;
• updating PromQL filters to use these variables (job="$job", instance=~"$instance").

Describe alternatives you've considered
The only alternative is to manually edit the JSON after import, which is error-prone and makes it difficult to reuse the dashboard across different environments or clusters.