Adding nodes without identifying bottleneck is ineffective

Increasing task nodes increases splits but does not improve query response time when the bottleneck is I/O rather than compute capacity. Changing configs randomly without understanding root cause provides no performance benefit.

Before scaling, use EXPLAIN ANALYZE to identify query bottlenecks. Export query JSON from Trino web UI for detailed analysis. Focus on understanding WHY queries are slow rather than randomly adjusting configurations or adding nodes.