Business Central should emit a built-in telemetry signal with aggregated usage statistics for database indexes and AL-defined keys across the entire environment, not only for ISV-owned applications.
Each AL secondary key creates and maintains a SQL index. Indexes improve read performance, but they also consume storage and add maintenance cost to insert, modify, delete, posting, import, synchronization, and upgrade scenarios. Today, partners and customers can analyze long-running SQL queries and table index lifecycle events, but there is no telemetry signal that periodically shows how each index is actually used.
Business Central 2026 wave 1 introduces Database Index Information in the client. Please extend this capability with telemetry: a periodic snapshot of index usage emitted to Application Insights.
The signal should include all index origins: Base Application, System Application, Microsoft apps, AppSource/ISV apps, per-tenant extensions, localizations, system-generated indexes, SIFT/VSIFT indexes, and protected primary/unique indexes.
The event should include table ID/name, company, AL key name and fields where available, SQL index name, index type, index origin, app ID/name/publisher/version, enabled status, index size, user seeks, scans, lookups, updates, last usage timestamps, read/write ratio, and a usage classification such as Unused, ReadHeavy, WriteHeavy, Balanced, or CandidateForReview.
The signal should be aggregated, for example, daily or weekly, not emitted on every index usage. Environment-level telemetry should include all indexes. Extension-level telemetry should include indexes owned by that extension so ISVs can optimize their own apps without exposing data about other publishers.
This would help customers, partners, and developers identify unused or write-heavy indexes, reduce storage usage, improve write performance, optimize AL key design, and make data-driven decisions when disabling indexes per company.
Comments
Great idea - I would really like to see this implemented. This kind of telemetry would make it much easier to make data-driven decisions, especially when identifying unused or costly indexes. It would be a real help in optimizing performance, storage costs, and AL key design.
Category: Development
