Snowflake is basically a cloud data warehouse. The basic unit of work in Snowflake is query. To monitor the load, we need to monitor the query load at a certain interval.
Two Available Options:
1. INFORMATION_SCHEMA Schema
2. ACCOUNT_USAGE Schema
Both these schema contains the QUERY_HISTORY view. Here the monitoring dashboard can be created by using either of the schema. But we need to understand the limitations of these schema, below are our findings
QUERY_HISTORY VIEW IN:
INFORMATION_SCHEMA: It contains only 10,000 records. In this case, one can't predict the result data belongs to the desired duration or not. As the 10K rows can be inserted as per the traffic at the time. In this case the INFORMATION_SCHEMA results can be inconsistent. It contains the data for 14 days interval OR 10K records.
ACCOUNT_USAGE : It contains the data for 1 year, with delay of 45 minutes.
Conclusion: In case, the required data analysis is continuous than only use the INFORMATION_SCHEMA.
In case, the required data analysis can accommodate delay of 45 minutes than the ACCOUNT_USAGE schema should be used.