Question

A logistics company stores GPS ping documents in an Azure Cosmos DB for NoSQL container that receives millions of writes per day. Each document contains latitude, longitude, driverId, and dozens of other telemetry fields. Analysts run occasional queries filtering only by driverId and a timestamp range. The team notices write RU charges are high and wants to reduce them without breaking the analytical queries. Reviewing the indexing policy, they find the default policy indexes all paths. What is the most effective change to reduce write RU consumption?

Accepted Answer

Set the includedPaths to only the driverId and timestamp paths and add the remaining paths (including the spatial fields) to excludedPaths.

Answer

Switch the indexing mode to 'none' so no paths are indexed, then rely on cross-partition scans for the analytical queries.

Answer

Add a spatial index on the latitude and longitude paths to speed up all queries and lower write cost.

Answer

Enable the integrated cache on a dedicated gateway to reduce the RU charge of the write operations.