Under Economy Scaling, a warehouse is started only if the system estimates there is enough query load to keep the warehouse busy for at least how many minutes?

Master the SnowPro Advanced Architect Test with flashcards, multiple-choice questions, and detailed explanations. Prepare thoroughly for your certification!

Multiple Choice

Under Economy Scaling, a warehouse is started only if the system estimates there is enough query load to keep the warehouse busy for at least how many minutes?

Explanation:
Economy Scaling focuses on cost-efficiently expanding warehouse capacity by avoiding unnecessary provisioning for short-lived bursts. The system weighs whether the expected query load will keep the warehouse busy long enough to justify the startup overhead and ongoing credits. In this approach, a new warehouse (or extra clusters within a multi-cluster warehouse) is only brought online if the predicted workload can sustain activity for a minimum period. That threshold is eight minutes, chosen to balance two goals: you want to capture meaningful bursts without paying for idle time, but you don’t want to delay provisioning so long that performance suffers during legitimate spikes. If the load isn’t expected to last that long, the system waits, avoiding the startup cost. If there is enough load for at least eight minutes, it scales up to handle the demand, improving performance and reducing queuing.

Economy Scaling focuses on cost-efficiently expanding warehouse capacity by avoiding unnecessary provisioning for short-lived bursts. The system weighs whether the expected query load will keep the warehouse busy long enough to justify the startup overhead and ongoing credits.

In this approach, a new warehouse (or extra clusters within a multi-cluster warehouse) is only brought online if the predicted workload can sustain activity for a minimum period. That threshold is eight minutes, chosen to balance two goals: you want to capture meaningful bursts without paying for idle time, but you don’t want to delay provisioning so long that performance suffers during legitimate spikes. If the load isn’t expected to last that long, the system waits, avoiding the startup cost. If there is enough load for at least eight minutes, it scales up to handle the demand, improving performance and reducing queuing.

Subscribe

Get the latest from Passetra

You can unsubscribe at any time. Read our privacy policy