Metrics

The Metrics dashboard at /metrics aggregates clone-operation KPIs across a rolling 7-day window. It's the operational dashboard for spotting throughput dips, success-rate regressions, and unusual activity patterns.

Core stats (top row)

Eight cards summarise the window:

Card	Meaning
Total clones	All jobs that started in window
Succeeded	Terminal succeeded count
Failed	Terminal failed count
Success rate	succeeded / (succeeded + failed)
Avg duration	Mean wall-clock per job
Min duration	Fastest job
Max duration	Slowest job
Tables cloned	Sum of tables_cloned across all jobs

Scale stats

A second row of four cards highlights aggregate scale:

• Data moved (sum of bytes transferred)
• Views cloned
• Avg tables per clone
• Week-over-week change — green/red badge with % delta

Charts

• 7-day activity — area chart of clones-per-day, hover for tooltip
• Status breakdown — donut split by succeeded / failed / partial
• Clone type — donut split by MANAGED / SHALLOW / DEEP
• Operation type — donut split by clone / sync / rollback
• Peak usage hours — bar chart of clones-by-hour-of-day across the window

Lists

• Top source catalogs — most-cloned-from catalogs with job count
• Active users — top users by job count

API

GET /metrics

Returns:

{
  "by_status": { "succeeded": 142, "failed": 8 },
  "clone_type_split": { "MANAGED": 130, "SHALLOW": 20 },
  "operation_type_split": { "clone": 110, "sync": 40 },
  "peak_hours": [{ "hour": 9, "count": 12 }, ...],
  "activity": [{ "date": "2026-04-25", "count": 28 }, ...],
  "wow": { "current": 150, "previous": 132, "delta_pct": 13.6 },
  "top_catalogs": [{ "catalog": "prod_warehouse", "count": 42 }],
  "active_users": [{ "user": "alice@acme.com", "count": 18 }]
}

The window length is configurable via the metrics_window_days setting in clxs.yaml (default 7).

Related

• Audit Trail — per-job drill-in
• Reports — exportable history
• Data Observability — health metrics for cloned tables