Snowflake Query Studio
Query Studio gives you visibility into the queries running against your Snowflake warehouses — historical and real-time — so you can find expensive or long-running queries, cancel a running query directly from ADOC, and get AI-assisted optimization advice.
Similar queries are grouped using fingerprinting: queries with the same structure share a fingerprint ID, giving you an aggregate view of how often a query pattern runs and what it costs, rather than reviewing every execution individually.
Snowflake Query Studio has three tabs: Query Studio, Query Insights, and Query Patterns (Beta).
Filters
Asset-level filters, in the left pane, plus the Global Calendar, control what data is shown across all three tabs.
Query Studio
Widget | Description |
|---|---|
Warehouse Workload Insights | A heat map showing the 95th-percentile value of Latency, Query Load, Queue Load, Blocked, and Credits for each warehouse, with above-normal values highlighted. Filterable by query status, warehouse, query type, database, and more — filters are contextual, so selecting a warehouse narrows the database filter to that warehouse's databases. Select refresh to fetch the latest data from Snowflake. |
Query Aggregate Table | Groups similar queries into a query fingerprint. You can also view queries individually with no fingerprint, or grouped by user. |
Query Fingerprint Table | Details for each fingerprint: Sample Query (select to open Query Details, or copy via the icon), Distinct Warehouses, Query Count, Cloud Services Credits, Query Cost, Avg/Max/Median Execution Time, Median Queued Provisioning Time, Avg Compilation Time, and Avg Total Elapsed Time. Filterable to top-50 categories like Top 50 Expensive Queries or Top 50 Long Running Queries — selecting a category updates the left-pane filters and limits results to that category. |
No Grouping Table | Every individual query, unfingerprinted: Query Text (copyable), Query Id, Execution Status, User Name, Query Cost, Credits Used Cloud Services, Warehouse Name (select to open Query Details, including actual cost incurred), Total Elapsed Time, Start/End Time, Bytes Scanned, Rows Produced, Warehouse Size, Query Type, Database Name, Schema Name, Execution Time, Compilation Time, Partitions Scanned, Partitions Total, % Partitions Scanned, Error Code, Error Message, Queued Provisioning/Repair/Overload Time, Transaction Blocked Time, and Bytes Spilled to Remote/Local Storage. |
Group By Users | Query activity grouped by user: User Name, Query Count, Avg/Max/Median Execution Time, Total Credits, and Query Cost. |
Select Export Raw Data to download the current view as a CSV, reflecting whatever filters are applied.
Real-time data
Turn on the Go Live toggle to view real-time Snowflake activity instead of historical data. You can set the refresh interval to every 1, 2, or 5 minutes. Filters can't be applied while Go Live is on.
Analyzing warehouse response time
Warehouse resizing is one of the simplest ways to control Snowflake cost, but it isn't a one-time task — warehouse cost, response time, and workload shift as the data flowing into your account changes, so resizing needs periodic review.
Establish a baseline response time for the warehouse, based on prior execution times under similar workload, warehouse size, query volume, and data handled.
Navigate to Query Studio.
Note the 95th-percentile execution time for queries on the warehouse, and compare it against your baseline to decide whether resizing is warranted.
Query Insights
Query and warehouse performance, broken down by execution stage (compilation time, execution time, and so on).
Widget | Description |
|---|---|
Queries Summary | Total Queries, Average Bytes Produced, Average Bytes Scanned, and Average Execution Time for the selected period. |
Query Time by Warehouse Name | Median execution timings (compilation, execution, and so on) per warehouse, plus a query-count trend line. Warehouses shown depend on your filters. |
Query Time Over Time | The same median timings, by time interval instead of by warehouse, based on the Global Calendar. |
Query Time by Database | The same median timings, by database. |
Query Time by Warehouse Size | The same median timings, by warehouse size (X-Small, Small, Medium, Large). |
Query Patterns (Beta)
Lets you go beyond a flat list of queries — such as Top 50 Expensive Queries — to see column-level patterns, including how often GROUP BY or table joins are used on specific columns. Use this to spot performance bottlenecks tied to particular columns rather than particular queries. Select a table to open its Details page for metadata and clustering information.
Query Advisor
Query Advisor is an AI-powered tool that analyzes a query and recommends specific optimizations, addressing common issues like expensive execution, long runtimes, high data scan volume, and departures from best practices.
To open Query Advisor:
Navigate to Compute and select a Snowflake data source.
Select Query Studio.
Select a query to open its Query Details page.
The Query Advisor section on that page shows:
Original Query Text — the full SQL query being analyzed.
Issue Overview — the specific inefficiencies found.
Optimization Recommendations — specific actions to improve efficiency, such as simplifying logic or reducing resource usage.
Expected Performance Improvement — an estimate of the gain from applying the recommendations.
Use the thumbs-up or thumbs-down feedback control to rate the analysis, with an optional comment, then select Submit.
What's next
Snowflake Performance – See how query activity relates to overall warehouse performance.
Snowflake Warehouse – Act on sizing recommendations informed by query load.
Send a comment