During our career as a database professional we likely encounter cases where a runaway query takes hours to complete or doesn't even complete by any reasonable time measurement. How do we troubleshoot such cases?
A query execution plan can help provide a conclusive explanation of query performance issues. But to get a query execution plan there is one requirement a long-running query can’t easily meet: query completion.
If a query takes a long time to complete or never actually does, then how can we troubleshoot those cases? And assume we take that production query back to our development server and it will run well? That means there is some set of conditions that can only be reproduced in the production server, be that the size of the database, the data distribution statistics, or even the availability of resources such as memory or CPU...