What to expect
With a typical setup of all ODP pods scaled to 1 instance each, we can expect the following numbers:
Insert (POST)
| Attributes/record | Payload size/request | Success rate | Throughput achieved |
23
| 1kb
| 100% | 40 tps
|
50
| 2kb
| 100%
| 35 tps |
120
| 5kb
| 100%
| 28 tps
|
Read (GET)
| Attributes/record | Payload size/request | Success rate | Throughput achieved |
23
| 1kb
| 100% | 85 tps |
50
| 2kb
| 100%
| 80 tps
|
120
| 5kb
| 100%
| 70 tps
|
Remove (DELETE)
| Attributes/record | Payload size/request | Success rate | Throughput achieved |
23
| 1kb
| 100% | 46 tps |
50
| 2kb
| 100%
| 41 tps
|
120
| 5kb
| 100%
| 37 tps
|
Key assumptions
- Tests performed on a 4-core 16 GB machine
- RAM utilization 50-70%
- CPU utilization 75-100%
- API hits include authentication and authorization
- No prehooks or posthooks as part of the data service
- History is enabled
- No relationships in the data service
