DynamoDB On-Demand Capacity Mode Transition and Billing Cycle Alignment

Capacity Mode Transition Billing Model Changes

AWS DynamoDB allows switching between provisioned and on-demand capacity modes, but the billing model changes immediately upon transition. When moving from provisioned to on-demand billing, the cost structure shifts from fixed capacity units to per-request pricing at 0.25 USD per million write request units and 0.13 USD per million read request units.

The transition timing affects billing cycles and may create partial-month cost calculations. AWS Console, CLI, or SDK updates trigger the billing model change, but the immediate cost impact depends on when during the billing period the change occurs.

Auto-scaling policies in provisioned mode use target utilization percentages to adjust capacity, while on-demand mode automatically scales without pre-configured limits. This fundamental difference affects how workloads are billed and optimized post-transition.

Key Questions

What is the exact timing mechanism AWS uses to apply the new billing model when transitioning from provisioned to on-demand capacity mode, and how does this affect partial-month cost calculations?

How can organizations verify the billing model change has taken effect and accurately predict cost impacts for workloads with unpredictable traffic patterns?

Are there documented differences in billing cycle alignment between provisioned and on-demand modes that could create unexpected cost spikes during transition periods?