避免 `nextRunAtMs` 前移导致 cron 漏跑

• Cron scheduler is enabled and `cron runs` history is queryable.
• You can create a temporary high-frequency job in non-production scope.

1. Create a test cron job with short interval (e.g., every 1-2 minutes) and fixed payload text.
2. Run scheduler under load and capture `nextRunAtMs` transitions in logs.
3. Compare expected trigger count vs `cron runs` history to detect skipped windows.
4. Add alerting when job delay exceeds threshold to catch future regressions early.

• Clock skew or host suspend can still affect schedule semantics and must be monitored separately.
• Detailed PR discussion may require authenticated GitHub view（需验证）。

This tip is aggregated from community/public sources and preserved with attribution.