Your DuckLake,
under control.
Observing the lake
Observing the lake
Built for data engineers
Features
Same Overview components and data shapes as the desktop app — hydrated here with sample lake stats.
lakemon — DuckLake production
247
12 schemas
3.2K
48600 MB total
892.1M
1.8K
847 eligible to expire
80
across 5 tables
threshold: 200/table
1.3K
487 older than 24h
cleanup threshold: 48h
847
eligible to expire
expire threshold: 45h
42.1K
pending flush
Files
Size (MB)
Rows
Avg KB
| Schema | Table | Files | Small | Size ↓ | Rows | Avg KB | |
|---|---|---|---|---|---|---|---|
| analytics | events | 847 | 23 | 12400.0 MB | 234.1M | 14950.4 KB | |
| warehouse | fact_orders | 156 | 45 | 8200.0 MB | 156.7M | 53760.0 KB | |
| analytics | user_sessions | 312 | 0 | 4800.0 MB | 89.2M | 15769.6 KB | |
| ml | training_data | 64 | 12 | 2100.0 MB | 45.3M | 33587.2 KB | |
| warehouse | dim_products | 28 | 0 | 128.0 MB | 1.2M | 4710.4 KB |
Lake Overview
Track total files, rows, writes, and deletions across snapshots. See how your lake evolves with interactive time-series charts — scoped to lake, schema, or table level.
Per-Table Detail
Click any table to see columns, types, partitions, live row count, churn percentage, and file distribution. Run targeted maintenance directly from the detail drawer.
Maintenance
Lakemon surfaces exactly which tables need maintenance and why — then lets you run it with a confirmation dialog showing the exact SQL that will execute.
Why Lakemon
DuckLake is powerful. Keeping it healthy shouldn't require a folder of ad-hoc queries and tribal knowledge.
-- TODO: find fragmented tables (run weekly)
SELECT schema_name, table_name,
count(*) as file_count
FROM ducklake_files
WHERE status = 'active'
GROUP BY 1, 2
HAVING count(*) > 50
-- and snapshots... where's that other script?
-- churn query is in Notion I think
Every. Single. Week.
Instant. Every time you open it.
Write SQL per table
Auto-flagged small files
Snapshots silently grow
One-click expire
Scattered notebooks
Single dashboard
Guess what needs compaction
Impact-ranked candidates
Row churn invisible
Per-table churn %
| Manual SQL | Lakemon | |
|---|---|---|
| Small files detection | Write queries per table | Auto-scanned, flagged per table |
| Snapshot hygiene | Forgotten, keeps growing | Eligible counts + one-click expire |
| Lake health overview | Scattered SQL scripts | Single dashboard view |
| Maintenance execution | Write SQL by hand | Preview SQL, confirm, run |
| File compaction | Guess which tables | Candidates ranked by impact |
| Row churn tracking | Not visible | Per-table churn % with alerts |
How it works
Point Lakemon at your DuckLake. Read-only by default — no data is ever modified without your explicit confirmation.
See schema stats, table health, small files, deletion queues, and snapshots in a single dashboard. Drill into any table.
Run compaction, rewrite, expire, and cleanup operations with one click. Preview the exact SQL before execution.
FAQ
Free and open source. Connects to any DuckLake instance via DuckDB.
macOS · Linux · Windows