In #1520 we hard-coded SORTABLE_DATASETS = new Set(["spans", "errors", "discover"]) based on live-org probe + existing spans guard.
We should have a canonical, machine-readable source (or at least a documented contract) rather than a static list that can drift. Possible homes:
- @sentry/api SDK
- Sentry server (events endpoint capability)
- Snuba schema / Discover docs
This would let the CLI (and other clients) query "which datasets accept sort on grouped aggregates?" instead of maintaining a hand-curated set.
cc @BYK
In #1520 we hard-coded
SORTABLE_DATASETS = new Set(["spans", "errors", "discover"])based on live-org probe + existing spans guard.We should have a canonical, machine-readable source (or at least a documented contract) rather than a static list that can drift. Possible homes:
This would let the CLI (and other clients) query "which datasets accept sort on grouped aggregates?" instead of maintaining a hand-curated set.
cc @BYK