FAQ

What fails if I omit dataset after December 3 2026?

Endpoints that default to discover are included in the retirement. After the cutoff, transactions/discover/events datasets are unavailable through the API, so omitted dataset behaves like discover and fails. Set dataset=spans or dataset=errors explicitly.

Why must I add is_transaction:true on spans queries?

Without it you query all spans, not only transaction-shaped ones. The old event.type:transaction filter maps to is_transaction:true.

How do I migrate percentile and apdex functions?

Use p50/p75/p90/p95/p99/p100 on span.duration instead of percentile(transaction.duration,…). Apdex and user_misery take (span.duration, threshold) with your own threshold.

What do I do if my old discover query mixed errors and transactions?

Split into one dataset=errors request and one dataset=spans request, then combine results in your application. A single discover call will not survive the cutoff.

Do SDK or Sentry UI need changes for this cutoff?

No. The change affects direct API requests only. SDKs keep sending data the same way; UI continues to work.

Why might counts differ after moving to spans?

The spans dataset adjusts counts for sampling to estimate total volume. Legacy transactions counted stored events. For raw sampled timeseries values you can set disableAggregateExtrapolation=1 where supported.

Is this official Sentry documentation?

No. Unofficial local helper. Not affiliated with Sentry.

Do you store my answers or auth tokens?

No. Local browser only; we never ask for auth tokens.