Analytics Warehouse Query Contention

A critical Scheduled Reports job with a real business deadline (e.g. month-end close, SLA-bound external reporting) is contending with ad-hoc BI/Data Science query load for the same shared Warehouse. Four business invariants govern every decision here: (1) critical scheduled-report jobs with real deadlines must complete within their SLA; (2) ad-hoc query access for BI/Data Science must not be revoked entirely — throttling, queueing, or prioritizing is the correct response, not an overcorrection that cuts off access; (3) killing a long-running Data Science job without warning has a real lost-work cost that must be weighed, never treated as free; (4) warehouse capacity alone is not the primary fix — prioritization and bounded throttling are. Decide how to prioritize the critical job over ad-hoc load, whether and how to throttle or gracefully stop long-running ad-hoc queries, and whether to communicate the tradeoff to the stakeholders it affects rather than silently starving them.

Difficulty

Senior Staff

Expected Time

~18 min

Category

Data Pipelines