Snowflake Query History Extractor
Extracts query history and performance metadata from Snowflake using the Snowflake Python Connector and INFORMATION_SCHEMA.QUERY_HISTORY view. Identifies expensive queries by credits consumed, data scanned, and spillage to remote storage. Exports results to a Pandas DataFrame for downstream analysis or Slack reporting.
npx skills add agentskillexchange/skills --skill snowflake-query-history-extractor
This skill connects to Snowflake using the snowflake-connector-python library with key-pair authentication or username/password. It queries the SNOWFLAKE.ACCOUNT_USAGE.QUERY_HISTORY view (or INFORMATION_SCHEMA.QUERY_HISTORY for session-scoped analysis) to retrieve query executions over a configurable time window. Queries are ranked by CREDITS_USED_CLOUD_SERVICES, BYTES_SCANNED, and BYTES_SPILLED_TO_REMOTE_STORAGE. The skill identifies warehouse sizing inefficiencies by comparing actual bytes scanned against partition pruning ratios derived from PARTITIONS_TOTAL and PARTITIONS_SCANNED columns. Results are loaded into a Pandas DataFrame and can be exported as CSV, pushed to a Snowflake staging table, or formatted as a Slack Block Kit message via the Slack Web API. Supports multi-account analysis using Snowflake Organization Account Usage views.