## Problem / Use case

You need to track datasets that undergo frequent schema updates. By focusing on `INFO`-level schema events and counting unique `snapshotId` values, this recipe helps highlight which datasets are changing most often — an important indicator for compliance and schema governance.

## Query

```dataprime
source system/engine.schema_fields
| filter $m.severity == INFO
| groupby dataset
    aggregate distinct_count(snapshotId) as schema_change_count
| sortby schema_change_count desc
```

## Expected output

| dataset              | schema_change_count |
| -------------------- | ------------------- |
| aaa.audit_events     | 10                  |
| engine.schema_fields | 10                  |
| logs                 | 10                  |
| labs.limitViolations | 10                  |
| spans                | 10                  |
| engine.queries       | 6                   |

## Variations

- Filter for `ERROR`-level events to identify failed or invalid schema updates.
- Add `max($m.timestamp)` to include the most recent change time per dataset.
- Combine with `count()` to measure total schema events alongside distinct snapshots.

## TL;DR

Count distinct schema snapshot IDs by dataset to surface frequently changing schemas — essential for monitoring data stability and compliance.