The dashboard interface
The dashboard is the main window into Log Lens: a single-page log dashboard and error tracker where you browse indexed issues, filter them, and inspect the events behind each one. Everything runs against your local SQLite index, so it stays fast and private.
Layout
The screen splits into three regions:
- Sidebar - pick the active application (exactly one; switching replaces the whole data context), browse indexed Sources, and read the Level distribution bars. You can collapse it and load more sources incrementally. (The API reference now lives in these docs, not as an in-app view — see API conventions.)
- Header - global search (
Cmd K), New issue, Tags, font scaling (90%-120%), theme toggle, settings, API key, refresh, Export CSV, and Process logs. - Overview - five compact headline widgets, including triage health, with daily activity embedded transparently beside the page title.
- Issue table - one row per normalized issue group, showing level, last observed, message, source, status, and lifetime count.
Operational overview
Operational context does not consume a separate dashboard row:
- Activity - separate red Error, orange Warning, and neutral Other trends with hollow points sit beside Error operations / Overview. Toggle each series with its pill, switch between 7, 14, and 30 days, hover for exact counts, and click a day to apply or clear the dashboard date filter. At least one series always remains visible.
- Triage health - the fifth headline widget shows the resolved percentage and workflow composition. Hover a segment for its status and count; click it to apply or clear the issue-table status filter.
Connector progress stays in Settings → Connectors, where it shows queued/running work, completed and confirmed actionable files, fetched bytes, the current filename, and recent outcomes. Unchanged remote files do not inflate the run total. While a run is active the dashboard polls every two seconds; closing settings or reloading the page does not affect the worker.
Filters and search
Search matches titles, messages, exceptions, frames, channels, and source paths. Filter by severity pill (All, CRITICAL, ERROR, WARNING, NOTICE, INFO, DEBUG), plus date, log type, source, status, tag, issue kind, and module. Sort by occurrence count, newest, or oldest, and page through 25/50/100/200 rows (default 50).
Every choice - view, search, filters, sort, page, and the open issue - is written to the URL, so a link reloads the exact same investigation.
Inspector
Clicking a row opens a 60%-width overlay (the table stays put) with four tabs:
| Tab | Contents |
|---|---|
| Overview | Daily occurrence trend chart, workflow status/notes/history, tag assignment |
| Stack trace | Parsed frames with Hide vendor and Clean paths toggles |
| Context | Structured context and the representative message |
| Occurrences | Occurrence metadata and links to the raw event |
Timeline and themes
The issue inspector’s Overview trend chart still plots occurrences for the selected issue. The main dashboard’s activity chart is application-wide. The theme button cycles system to light to dark; the preference is stored in your browser and applied before render to avoid a flash.
Settings tabs
The settings action opens a tabbed panel:
- Connectors - create, test, sync, or delete local-directory and SSH sources.
- Ingestion - choose which severity levels future imports retain.
- Maintenance - scoped deletion and processed-archive retention.