Sentry vs GlitchTip vs Self-Hosted Sentry: Error Tracking for a Small Team When the Event Quota Keeps Running Out

Описание и применение

Для небольших команд Sentry может быстро исчерпать квоту из‑за шумных ошибок. Добавление клиентского дедупликации через beforeSend, ignoreErrors и sampleRate снижает количество событий, предотвращая переплату.

Задача

Small teams run out of event quota on hosted Sentry due to noisy incidents, making error tracking costly.

Результат

Implementing client‑side deduplication and sampling reduces event volume by an order of magnitude, keeping hosted Sentry within quota and avoiding overage costs.

Рабочий процесс

• 1. Identify noisy incidents that generate repeated errors.
• 2. Add ignoreErrors for known network noise (e.g., ECONNRESET, AbortError).
• 3. Implement a beforeSend limiter that tracks error fingerprints and drops events exceeding a per‑window threshold.
• 4. Configure sampleRate for high‑volume noise if needed.
• 5. Verify that events are dropped before reaching the Sentry server.
• 6. Optionally enable server‑side spike protection or inbound filters for additional safety.

Доказательства

Drop repeats at the source. The Sentry JavaScript/Node SDK exposes beforeSend, ignoreErrors and sampleRate

Источник

A ten‑line beforeSend limiter removes the failure mode that makes every error‑tracking plan look too small.

Источник