Question

You maintain a Cosmos DB container for an e-commerce platform that stores orders, partitioned by /customerId. A reporting feature must return, for a single specified customer, the total amount spent per month. The query currently uses GROUP BY on month and executes across all partitions, consuming more RUs than expected. Which change will most effectively reduce the RU cost of this specific query?

Accepted Answer

Add a WHERE clause filtering on customerId so the query targets a single logical partition.

Answer

Switch the container's partition key to /month to co-locate all orders by month.

Answer

Move the GROUP BY logic into a stored procedure so aggregation runs server-side.

Answer

Enable autoscale throughput so the query has more RUs available during peak reporting.