snowflake_billing_warehouse_credits_used

The snowflake_billing_warehouse_credits_used metric measures the cumulative number of Snowflake credits consumed by virtual warehouses for compute operations. As a counter-type metric, it continuously increments to track the total credit expenditure across warehouse usage, which directly translates to cloud infrastructure costs. This metric is fundamental to understanding Snowflake's consumption-based pricing model, where virtual warehouses consume credits based on their size and the time they remain active. Each credit represents a specific amount of computational resources, and monitoring this metric provides visibility into the primary cost driver for most Snowflake deployments. Organizations can track credits used per warehouse, enabling granular cost attribution across different workloads, teams, or applications.

From an operational perspective, this metric is critical for cost management, capacity planning, and identifying inefficiencies in warehouse utilization. By analyzing credit consumption patterns, teams can detect anomalies such as warehouses left running idle, oversized warehouses for light workloads, or unexpected spikes in query processing that drive up costs. The metric should be monitored alongside warehouse uptime and query patterns to calculate effective cost per query or cost per hour. According to Snowflake's resource monitoring documentation, establishing resource monitors with credit quotas can help prevent budget overruns. Healthy patterns typically show predictable credit consumption aligned with business hours and known workload schedules, with minimal credit usage during non-peak periods when automated warehouse suspension should be in effect.

Common alerting use cases include setting thresholds for unexpected credit consumption rates (credits per hour exceeding baseline), detecting warehouses consuming disproportionate credits relative to their expected workload, and triggering notifications when approaching monthly or quarterly credit budgets. For troubleshooting, sudden increases in this metric often indicate runaway queries, missing warehouse auto-suspend configurations, or newly deployed workloads requiring optimization. Teams should correlate spikes with the WAREHOUSE_METERING_HISTORY view in Snowflake's Account Usage schema to identify specific warehouses and time periods driving cost increases, enabling targeted investigation and remediation.