Firestore automatically creates single field indexes for most fields which means each document write updates multiple index entries and drives up write and storage costs. By trimming indexes to only the ones required by your queries on steps, sessionDuration and energyBurned you reduce index write amplification and index storage while preserving the performance of the queries you care about. This is a configuration change that does not reduce availability because Firestore remains fully managed and you keep the essential indexes for your filters and sort orders.
You can delete unused composite indexes and add single field index exemptions for attributes that are not part of your supported query patterns. Monitor index usage and keep only the minimal set that powers production queries so you sustain acceptable latency while lowering cost.