Traffic is steady and high, and you need consistent throughput on Bedrock. What fits?
Why A is the answer
Provisioned Throughput
Provisioned Throughput reserves capacity for sustained workloads.
In detail
- Provisioned Throughput gives consistent capacity for steady traffic.
- Commit only after measuring steady demand.
Use case
An illustrative scenario, not a real incident.
A support app has steady high volume, and Provisioned Throughput reserves the capacity it needs.
See it in different ways
flowchart TD
A{"What do you need?"} -->|"Variable traffic, managed models"| B["Bedrock on-demand"]
A -->|"Consistent high throughput"| C["Bedrock Provisioned Throughput"]
A -->|"Your own model container"| D["SageMaker AI endpoint"]
A -->|"Offline volume"| E["Batch inference"]mindmap
root((Inference))
On demand
Variable traffic
Provisioned Throughput
Steady and high
Batch
Offline and cheaper
Cross-Region
Spreads peaks
Streaming
Show text as it arrivesflowchart TD
Q{"What is the traffic like?"}
Q -->|"Steady and high"| A0["Provisioned Throughput"]
Q -->|"Variable"| A1["On demand"]Key point: Provisioned Throughput suits sustained workloads.
Why the other options are wrong
- B. On-demand use with a much longer request timeout
- A timeout does not add capacity.
- C. Batch inference that runs once per day for all requests
- A daily batch is not live.
- D. A larger Lambda function memory size for the application
- Lambda memory is not model capacity.