10 Mar 20 · karthikv · #Snowflake ·   Bookmark   ×

Undersized Warehouse/cluster that are incurring too much query time in Snowflake

How to identify undersized virtual warehouse or cluster in snowflake.

karthikv

posted on 10 Mar 20

Enjoy great content like this and a lot more !

Signup for a free account to write a post / comment / upvote posts. Its simple and takes less than 5 seconds




1. Go through the QUERY_HISTORY view and check EXECUTION_TIME to find out the list of top time consuming queries

2. Now, lets check, if the query was too large for the warehouse size to handle. Check for metrics like PARTITIONS_SCANNED, BYTES_SPILLED_TO_LOCAL_STORAGE, BYTES_SPILLED_TO_REMOTE_STORAGE etc...

3. Also, look for the current warehouse load at that point in time and find out if the warehouse was overloaded / had queued queries by looking at WAREHOUSE_LOAD_HISTORY

We can combine all three steps into a single query, i would do that some point in the future. If you are able to achieve it, I would appreciate if you share the query

More reading:

Thanks npack. I am working as a snowflake administrator. Can you please guide me what are all the responsibilities should i need to take care.

Security, Backups and restores, Cost optimization, establishing best practices, helping integration with other tools are some of the top admin duties i can think of


Community Software by Hittly
Copied