Snowflake Billing and Cost Management

Snowflake billing has two axes: compute (credits) and storage (TB/month). Compute consumes credits based on warehouse uptime and serverless service usage, while storage is billed against the combined volume of active data, Time Travel, and Fail-safe. This article walks through the structure of the credit billing model, the SQL you need for monitoring, Resource Monitor configuration, and 5 cost-optimization tactics you can apply in production right away.

Credit Billing Model

Warehouse Size and Credit Consumption

For Multi-cluster Warehouses, you multiply the credits above by the number of running clusters. For example, a Medium warehouse with 3 active clusters consumes 4 x 3 = 12 credits/hour.

WAREHOUSE_METERING_HISTORY

-- Daily credit consumption per warehouse
SELECT
  start_time::DATE AS usage_date,
  warehouse_name,
  SUM(credits_used) AS total_credits,
  SUM(credits_used_compute) AS compute_credits,
  SUM(credits_used_cloud_services) AS cloud_credits
FROM SNOWFLAKE.ACCOUNT_USAGE.WAREHOUSE_METERING_HISTORY
WHERE start_time >= DATEADD(MONTH, -1, CURRENT_TIMESTAMP())
GROUP BY 1, 2
ORDER BY total_credits DESC;

-- Hour-of-day credit consumption pattern
SELECT
  EXTRACT(HOUR FROM start_time) AS hour_of_day,
  warehouse_name,
  AVG(credits_used) AS avg_hourly_credits
FROM SNOWFLAKE.ACCOUNT_USAGE.WAREHOUSE_METERING_HISTORY
WHERE start_time >= DATEADD(DAY, -30, CURRENT_TIMESTAMP())
GROUP BY 1, 2
ORDER BY warehouse_name, hour_of_day;

METERING_HISTORY (All Services Combined)

-- Credit consumption across all services
SELECT
  start_time::DATE AS usage_date,
  service_type,
  SUM(credits_used) AS total_credits
FROM SNOWFLAKE.ACCOUNT_USAGE.METERING_HISTORY
WHERE start_time >= DATEADD(MONTH, -1, CURRENT_TIMESTAMP())
GROUP BY 1, 2
ORDER BY usage_date, total_credits DESC;

-- service_type values:
-- WAREHOUSE_METERING    : Warehouse
-- AUTO_CLUSTERING       : Automatic Clustering
-- MATERIALIZED_VIEW     : MV maintenance
-- PIPE                  : Snowpipe
-- REPLICATION           : Replication
-- SEARCH_OPTIMIZATION   : Search Optimization Service
-- SERVERLESS_TASK       : Serverless Task

Resource Monitor

Resource Monitor is the governance feature that caps warehouse credit consumption and triggers notifications or suspension when thresholds are hit. You create and manage them as ACCOUNTADMIN.

-- Create a Resource Monitor
CREATE RESOURCE MONITOR bi_monitor
  WITH CREDIT_QUOTA = 500
  FREQUENCY = MONTHLY
  START_TIMESTAMP = IMMEDIATELY
  TRIGGERS
    ON 80 PERCENT DO NOTIFY
    ON 95 PERCENT DO SUSPEND
    ON 100 PERCENT DO SUSPEND_IMMEDIATE;

-- Assign the Resource Monitor to a warehouse
ALTER WAREHOUSE bi_dashboard_wh SET RESOURCE_MONITOR = bi_monitor;

-- Account-level Resource Monitor (applies to every warehouse)
ALTER ACCOUNT SET RESOURCE_MONITOR = account_monitor;

-- Inspect Resource Monitors
SHOW RESOURCE MONITORS;

Resource Monitor Action Comparison

5 Cost Optimization Tactics

1. Tune AUTO_SUSPEND

-- Interactive: auto-suspend after 5 minutes (300 seconds)
ALTER WAREHOUSE bi_wh SET AUTO_SUSPEND = 300;

-- Batch: auto-suspend after 1 minute (60 seconds)
ALTER WAREHOUSE etl_wh SET AUTO_SUSPEND = 60;

-- Enable AUTO_RESUME so the warehouse wakes up on query arrival
ALTER WAREHOUSE bi_wh SET AUTO_RESUME = TRUE;

2. Separate Warehouses by Workload

Running BI, ETL, and data science workloads on a single warehouse causes resource contention, slows every query, and ultimately drives up cost. Create a dedicated warehouse per workload and tune size and AUTO_SUSPEND individually for each.

3. Reduce Scan Volume with Query Optimization

-- Find queries with poor partition pruning
SELECT
  query_id,
  query_text,
  partitions_scanned,
  partitions_total,
  ROUND(partitions_scanned / NULLIF(partitions_total, 0) * 100, 1) AS scan_pct,
  bytes_scanned / POWER(1024, 3) AS gb_scanned
FROM SNOWFLAKE.ACCOUNT_USAGE.QUERY_HISTORY
WHERE start_time >= DATEADD(DAY, -7, CURRENT_TIMESTAMP())
  AND partitions_total > 100
  AND partitions_scanned / NULLIF(partitions_total, 0) > 0.8
ORDER BY bytes_scanned DESC
LIMIT 20;

4. Cut Storage Cost

-- Inspect Time Travel and Fail-safe storage usage
SELECT
  table_catalog AS database_name,
  table_schema AS schema_name,
  table_name,
  ROUND(active_bytes / POWER(1024, 3), 2) AS active_gb,
  ROUND(time_travel_bytes / POWER(1024, 3), 2) AS time_travel_gb,
  ROUND(failsafe_bytes / POWER(1024, 3), 2) AS failsafe_gb,
  ROUND(retained_for_clone_bytes / POWER(1024, 3), 2) AS clone_gb
FROM SNOWFLAKE.ACCOUNT_USAGE.TABLE_STORAGE_METRICS
WHERE active_bytes > 0
ORDER BY (active_bytes + time_travel_bytes + failsafe_bytes) DESC
LIMIT 30;

-- Detect unnecessary Transient candidates (intermediate data that doesn't need Fail-safe)
-- Switching to Transient tables cuts Fail-safe storage cost

5. Monitor and Control Serverless Credits

-- Serverless credit trend by service
SELECT
  start_time::DATE AS usage_date,
  service_type,
  SUM(credits_used) AS credits
FROM SNOWFLAKE.ACCOUNT_USAGE.METERING_HISTORY
WHERE start_time >= DATEADD(MONTH, -1, CURRENT_TIMESTAMP())
  AND service_type != 'WAREHOUSE_METERING'
GROUP BY 1, 2
ORDER BY usage_date, credits DESC;

Chargeback Report

-- Monthly cost report per warehouse (for chargeback)
SELECT
  warehouse_name,
  DATE_TRUNC('MONTH', start_time) AS billing_month,
  SUM(credits_used) AS total_credits,
  SUM(credits_used) * 3.00 AS estimated_cost_usd  -- unit price depends on contract
FROM SNOWFLAKE.ACCOUNT_USAGE.WAREHOUSE_METERING_HISTORY
WHERE start_time >= DATEADD(MONTH, -3, CURRENT_TIMESTAMP())
GROUP BY 1, 2
ORDER BY billing_month DESC, total_credits DESC;

Check Yourself with a Question

Frequently Asked Questions