Limits & Performance
To keep Prism fast and reliable for everyone, interactive queries follow limits. Exact values can vary by plan.
Limits (Typical)
- Runtime: queries stop after a set duration if unfinished.
- Row/size: result sets are capped to a maximum row count and file size.
- Concurrency: a small number of queries can run at once per organization; additional runs queue.
- Auto-LIMIT: if no
LIMIT
is present, Prism may add one automatically.
Performance Tips
- Filter by date window and status; avoid scanning entire history.
- Select only needed columns; prefer explicit lists over
SELECT *
. - Aggregate early with
GROUP BY
andWHERE
predicates. - Preview with
LIMIT
before removing or increasing the cap. - Use schedules for heavy or recurring jobs instead of re-running interactively.
When a Limit Is Reached
- Timeouts: reduce the time range, simplify the query, or add indexes/filters if available.
- Large results: aggregate, reduce columns, or export in batches by date.
- Concurrency: wait for in-flight queries to finish or cancel one from your runs list.
Last updated: 2025-10-20
Updated about 2 hours ago
What’s Next