原文
Choose how big your data is and how much throughput you need. Each point is a benchmarked instance × disk; everything right of the target line clears your requirement, and the cheapest of those is the pick.
Want the raw numbers instead? See the full benchmark data →
Context
- Workload
- 90% read / 10% write
- Concurrency
- 32
- PostgreSQL
- 17
- Region
- us-east-1
- Pricing
- on-demand