The charm does not surface workload error logs

[sed-i]

Bug Description

When the workload emits error level logs, the only way to see them is to kubectl logs, or rely on pebble log forwarding + loki.
The charm is active/idle while in practice we have "model departure": the dashboards forwarded over reldata are not present in grafana's database.

To Reproduce

Forward an invalid dashboard to grafana, e.g. with a too long UID.

Environment

Any

Relevant log output

2024-12-13T20:17:08.624Z [grafana] logger=provisioning.dashboard type=file name=Default t=2024-12-13T20:17:08.623968857Z level=error msg="failed to save dashboard" file=/etc/grafana/provisioning/dashboards/juju_grafana-agent_f70bf56.json error="uid too long, max 40 characters"
2024-12-13T20:17:13.530Z [grafana] logger=provisioning.dashboard type=file name=Default t=2024-12-13T20:17:13.530628028Z level=error msg="failed to save dashboard" file=/etc/grafana/provisioning/dashboards/juju_grafana-agent_f70bf56.json error="uid too long, max 40 characters"
2024-12-13T20:17:18.535Z [grafana] logger=provisioning.dashboard type=file name=Default t=2024-12-13T20:17:18.535278997Z level=error msg="failed to save dashboard" file=/etc/grafana/provisioning/dashboards/juju_grafana-agent_f70bf56.json error="uid too long, max 40 characters

Additional context

No response