AEP Ad Hoc Query — 10-minute timeout

The rule. Ad Hoc Queries against the AEP Data Lake terminate at 10 minutes (source.packaged-vs-composable-md).

Why it's a problem. In a composable ecosystem, data engineers routinely run complex ELT jobs that join massive fact tables (billions of web events) with large dimension tables. A 10-minute cap forces engineers into awkward workarounds:

• Micro-batching — breaking queries into highly specific time windows.
• Over-reliance on Query Service CTAS — being forced to schedule Create-Table-As-Select operations just to get around ad-hoc limits, which clutters the data lake with intermediary datasets.
• Stifled exploration — data scientists and analysts cannot easily explore unstructured or raw data organically; the fear of timing out looms over every complex JOIN or window function (source.packaged-vs-composable-md).

Composable counterpart. Cloud data warehouses scale compute elastically; complex queries that take 30+ minutes complete without intervention, gated only by cost (tradeoff.query-flexibility).