Explanation
Tables experience frequent and large-scale modifications ("high churn"), causing Snowflake to constantly recluster data. This results in substantial compute consumption for maintenance tasks. Contributing factors include suboptimal clustering key selections, unordered data loads, and frequent full-table replacements.
Relevant Billing Model
Auto-Clustering consumption is billed as Snowflake compute credits. Costs increase based on the volume of data reclustered and the frequency of clustering operations, regardless of query activity.
Detection
- Review Auto-Clustering activity metrics for tables with >30-40% monthly reclustering
- Analyze loading patterns for high-churn behaviors
- Evaluate whether clustering keys align with actual query patterns
- Assess table criticality
Remediation
- Optimize data loading through incremental approaches and pre-sorting
- Redesign cluster keys to prioritize commonly filtered/joined columns
- Disable clustering for low-value tables
- Monitor costs against evolving access patterns