Feature Description
The current Logtide dashboard is fixed and organization-wide. This issue introduces user-configurable dashboards where teams can build the view that fits their stack, pinning the panels that matter most to them.
Problem/Use Case
Different teams within the same organization have very different observability needs. A backend team cares about p99 latency, error rates, and slow DB queries. A security team wants detection events, open incidents, and top triggered Sigma rules front and center. Right now everyone sees the same fixed dashboard, which means no one sees exactly what they need without navigating across multiple sections.
Proposed Solution
- Dashboard builder UI: add, remove, and resize panels — ideally with drag-and-drop
- Panel types available: time-series chart, single stat card, top-N table, live log stream, alert status indicator
- Data sources for panels: metrics (once #1 is implemented), log volume aggregations by level/service, span latency from traces, detection events from SIEM
- Per-project dashboards (team-specific views) and an optional global cross-project view
- Dashboard configurations stored as JSON in the database, with optional YAML export for version-controlling dashboards alongside infrastructure code
Alternatives Considered
- Grafana embed: embedding Grafana panels inside Logtide. Rejected because it introduces a heavy external dependency, defeats the self-contained deployment model, and complicates the privacy/GDPR positioning.
- Fixed dashboard with more widgets: simply adding more panels to the current static dashboard. Rejected because it doesn't scale to different team needs and creates a cluttered UI.
Implementation Details (Optional)
- Panel queries must target existing continuous aggregates (logs_hourly_stats, spans_hourly_stats, detection_events_hourly_stats) to maintain the <100ms render performance already established
- Dashboard JSON schema should be versioned from day one to allow future migrations without breaking saved configurations
- Consider a library like
svelte-dnd-action for drag-and-drop panel reordering in the SvelteKit frontend
Priority
Target Users
Additional Context
This feature also serves as the natural home for metrics visualizations once OTLP metrics ingestion is available (see issue #1). Building the dashboard infrastructure first means metrics panels can be added as a new data source type without redesigning the UI.
Contribution
Feature Description
The current Logtide dashboard is fixed and organization-wide. This issue introduces user-configurable dashboards where teams can build the view that fits their stack, pinning the panels that matter most to them.
Problem/Use Case
Different teams within the same organization have very different observability needs. A backend team cares about p99 latency, error rates, and slow DB queries. A security team wants detection events, open incidents, and top triggered Sigma rules front and center. Right now everyone sees the same fixed dashboard, which means no one sees exactly what they need without navigating across multiple sections.
Proposed Solution
Alternatives Considered
Implementation Details (Optional)
svelte-dnd-actionfor drag-and-drop panel reordering in the SvelteKit frontendPriority
Target Users
Additional Context
This feature also serves as the natural home for metrics visualizations once OTLP metrics ingestion is available (see issue #1). Building the dashboard infrastructure first means metrics panels can be added as a new data source type without redesigning the UI.
Contribution