Detector fails on folder dashboards and dashboards with missing data

[FF186]

Description:
While running the Angular Detector against a Grafana instance, I noticed that some dashboards fail to download, causing the run to terminate with errors. There appear to be two main issues:

1. Folders are treated as dashboards

   • The current code calls GetDashboard on all entries, including folders, which causes 404 errors for folders.

2. Dashboards with missing or empty data

   • Some dashboards return 406 errors when calling GetDashboard.
   • This seems to happen for dashboards that exist but either have missing panels or are inaccessible via the API due to format, permissions, or missing content.
   • These dashboards fail the run completely instead of being skipped or logged, making it hard to process large instances.

Steps to reproduce:

1. Run the Angular Detector against a Grafana instance containing a mix of dashboards and folders.

2. Observe errors like:

   get dashboard "dasdh1rcs5q8d": bad status code: 404
   get dashboard "wjbtmtb": bad status code: 406
   

3. Inspect dash.Type and note that 404 dashboards are actually folders.

Expected behavior:

• The detector should skip folders automatically.
• Dashboards with missing data or inaccessible via API should log a warning but not abort the entire run.
• Final output should include all successfully processed dashboards and a list of skipped/errored dashboards.

Suggested fix / possible approach:

• Check dash.Type before calling GetDashboard and skip folders.
• Add more verbose logging for dashboards that return 406 errors (UID, URL, Type) to help identify the reason.
• Consider collecting download errors separately rather than aborting the entire run.

Environment:

• Grafana version: Grafana v12.3.1 (3a1c80ca7c)
• Detector commit: sha256:fff0c3435efe267bb4442493f4610d8de53f996008d92b735399ed320b8a1d6f
• Go version: go version go1.26.0 linux/amd64