Scaling and Performance
Orbit is designed for scalable deployments and high performance:
- Docker Compose: Supports multi-stage builds for efficient development and production images.
- Postgres: Exposed on
localhost:5435
for local development; persisted in host-mounted volumes. - Backups: Use
pg_dump
andpg_restore
for database backup and recovery. - Vector Extension: Enable
pg_vector
for AI workloads.
Example: Database Backup
pg_dump --host localhost --port 5435 --user postgres --dbname ai_poc --format=d --jobs=2 --file ~/Downloads/ai_poc_dump
Example: Database Restore
pg_restore --host localhost --port 5435 --user postgres --dbname ai_poc --format=d --jobs=2 ~/Downloads/ai_poc_dump
Performance Tuning
Monitor with Grafana and Prometheus. Adjust Docker and Postgres settings as needed for your environment.
Dashboard Refresh Handling
Orbit dashboards can be refreshed in three main ways to keep your data up to date:
1. Manual Refresh Button
- Users can click a refresh button in the dashboard to instantly update all data.
- The dashboard shows a loading state while refreshing.
- All dashboard components are updated at once.
- There is a default 10 minute cooldown period between refreshes.
2. Automatic Refresh on Login
- When a user logs in, their most recently viewed dashboards are refreshed automatically in the background.
- This ensures you see the latest data without waiting.
- Only dashboards accessed in the last week are refreshed to reduce unnecessary load.
3. Scheduled Background Refresh
- The system can periodically refresh dashboards for all users or organizations, even if no one is actively using them.
- This keeps data fresh and reliable for everyone.
- Refreshes can be filtered by organization or date.
Summary Table
Method | Trigger | Scope | Parallelism |
---|---|---|---|
Manual Refresh Button | User clicks button | Current dashboard | All components |
Automatic Login Refresh | User logs in | Recent dashboards | Multiple dashboards |
Scheduled Background Job | System schedule | Org/all dashboards | Multiple dashboards |
Notes
- All refresh methods use the same core logic for updating dashboard data.
- Dashboards are refreshed in batches for speed, but components within a dashboard are updated one after another for consistency.
- The system includes retry and error handling to ensure reliability.
Last updated on