Question

On SQL Server, the orders DB starts throwing widespread lock-timeout errors at 02:10 and writes back up across many unrelated tables, not just one. Dashboards: blocking-chain length spikes, and the blocking-tree root is a single session running a long analytics report that began at 02:00. That session holds escalated table-level (not row-level) locks on `orders` and `order_items`, blocking all writers behind it. A nightly reporting job was rescheduled this week to run during a busier window. Triage and explain the cascade.