search
Go to website

slidersEfficiency Indicators

A detailed look at each efficiency indicator: what it measures, how the analysis works, and what actions you can take.

The Snowflake Efficiency Score evaluates four distinct areas of warehouse and compute efficiency. Each indicator maps to a specific optimization opportunity within Seemore.

hashtag
Warehouse Idle Time

Lookback window: 21 days Related feature: Auto-Shutdown

hashtag
What It Measures

Warehouse Idle Time tracks periods when a warehouse is running (consuming credits) but no queries are being executed. This is one of the most common sources of unnecessary Snowflake spend.

hashtag
How the Analysis Works

  1. The app queries ACCOUNT_USAGE.WAREHOUSE_EVENTS_HISTORY to identify warehouse start/stop events and query execution windows.

  2. It calculates the gap between the end of one query and the start of the next for each warehouse.

  3. Credits consumed during these idle gaps are aggregated and compared to total credits consumed.

hashtag
What You Can Do

  • Enable Auto-Shutdown in Seemore to automatically suspend warehouses the moment they become idle, without interrupting active queries.

  • Review the per-warehouse breakdown in the results table to identify which warehouses have the highest idle cost.

hashtag
Multi-Cluster Idle Time

Lookback window: 7 days Related feature: Auto-Scaler Requires: Snowflake Enterprise edition or higher

hashtag
What It Measures

For warehouses configured with multi-cluster scaling, this indicator measures idle time on non-primary clusters — additional clusters that were scaled up but had no workload to process.

hashtag
How the Analysis Works

  1. The app queries ACCOUNT_USAGE.WAREHOUSE_EVENTS_HISTORY filtered to non-primary clusters (cluster number > 1).

  2. It identifies periods where additional clusters were running but not serving queries.

  3. The idle cost is calculated based on credits consumed by those clusters during idle periods.

hashtag
What You Can Do

  • Review your multi-cluster scaling policies — consider tighter scaling thresholds or shorter scale-down windows.

  • Use Seemore's Auto-Scaler recommendations to optimize when additional clusters spin up and down.

Note: If your account uses Standard edition, this indicator is skipped and the overall score is calculated from the remaining three indicators.

hashtag
Warehouse Sizing Efficiency

Lookback window: 7 days Related feature: Smart Pulse

hashtag
What It Measures

Sizing Efficiency evaluates whether each warehouse is appropriately sized for its workload. An oversized warehouse wastes credits per query; an undersized one leads to longer execution times and queue delays.

hashtag
How the Analysis Works

  1. The app queries ACCOUNT_USAGE.QUERY_HISTORY to retrieve execution times and warehouse sizes for recent queries.

  2. It simulates how queries would perform at one size smaller and one size larger.

  3. A sizing efficiency percentage is computed based on how many warehouses are at or near their optimal size.

hashtag
What You Can Do

  • Review the sizing recommendations in the results table to see which warehouses are candidates for resizing.

  • Enable Smart Pulse in Seemore to automatically adjust warehouse sizes based on workload patterns throughout the day.

hashtag
Clustering Column Efficiency

Lookback window: 7 days Related feature: Auto-Clustering Agent

hashtag
What It Measures

Clustering Column Efficiency evaluates how effectively your tables' clustering keys are configured. Poor clustering leads to excessive partition scanning, increasing query cost and execution time. Unnecessary auto-clustering incurs maintenance costs without meaningful performance gains.

hashtag
How the Analysis Works

  1. The app queries ACCOUNT_USAGE.AUTOMATIC_CLUSTERING_HISTORY to identify tables with active auto-clustering and their associated costs.

  2. It queries ACCOUNT_USAGE.TABLE_PRUNING_HISTORY to measure partition pruning effectiveness — the ratio of partitions scanned versus partitions pruned.

  3. Tables with low pruning ratios or high clustering costs relative to benefit are flagged.

hashtag
What You Can Do

  • Review the per-table clustering analysis to identify tables where clustering keys should be changed, added, or removed.

  • Use Seemore's Auto-Clustering Agent to get AI-driven recommendations for optimal clustering key selection, including cost-benefit analysis for each table.

PreviousUnderstanding Your ScoreNextSettings

Last updated