metrics-pipeline

A single-writer metrics pipeline coordinates MySQL and Redis updates by applying a four-state processing model with fingerprint-based idempotency, ensuring all photo events produce consistent, auditable analytics.

Use Redis as a derived cache for leaderboards while maintaining the authoritative metrics in MySQL. This enables safe rebuilds from the metrics table and ensures deterministic upserts across multiple timescales and scopes.

Prevent duplicate calculations using fingerprint-based idempotency. The pipeline tracks four fields—processed_at, processed_fp, processed_tags, and processed_xp—to ensure identical photo events do not trigger redundant metric updates.

Yes, the metrics pipeline supports per-user XP processing and tracking. It centralizes XP computations alongside tag verifications and admin actions to feed accurate data into Redis leaderboards and operational dashboards.

Safely rebuild Redis caches from the MySQL metrics table. Because Redis is a derived cache and the single-writer MetricsService governs all state, you can repopulate leaderboards and per-user XP deterministically without data loss.

Trigger metrics pipeline updates for all photo processing events including tags, verification, admin actions, and deletion. The MetricsService processes these events to maintain dashboards, leaderboards, and XP computations accurately.