How latency is calculated in the caliper?
Is latency calculated only for the successful transactions? Why latency is not calculated for all the transactions?
Context
I am sending 2000 transaction with send rate varying from 5tps-150tps. I am using sample ethereum network. After a certain send rate, many of the transactions are failed. And the system is able to serve certain number of transactions as given by succ in the table below. The latency is expected to increase exponentially but it remains constant.
I wanted to understand, is the latency calculated only for the successful transaction. Why the latency is not considered for all (succ + fail) transactions?
| input tps | Succ | Fail | Send Rate (TPS) | Max Latency (s) | Min Latency (s) | Avg Latency (s) | Throughput (TPS) |
|---|---|---|---|---|---|---|---|
| 5 | 2000 | 0 | 5 | 1.18 | 1.01 | 1.07 | 5 |
| 10 | 2000 | 0 | 10.1 | 13.16 | 1.03 | 10.08 | 9.8 |
| 15 | 1151 | 849 | 15.2 | 50.92 | 1.06 | 28.65 | 11 |
| 20 | 770 | 1230 | 20.3 | 51.03 | 1.04 | 27.15 | 13.4 |
| 30 | 603 | 1397 | 30.5 | 51.08 | 1.06 | 27.16 | 17.2 |
| 40 | 541 | 1459 | 41.7 | 50.75 | 1.8 | 27.19 | 20.4 |
| 50 | 515 | 1485 | 52.5 | 50.85 | 1.79 | 27.21 | 22.5 |
| 60 | 504 | 1496 | 62.1 | 50.97 | 1.05 | 26.88 | 24.1 |
| 70 | |||||||
| 80 | 478 | 1522 | 85.9 | 51.11 | 1.85 | 27.28 | 27 |
| 90 | |||||||
| 100 | 465 | 1535 | 105.1 | 50.63 | 1.07 | 26.85 | 28.6 |
| 125 | 460 | 1540 | 134.9 | 50.71 | 1.24 | 26.78 | 30.8 |
| 150 | 451 | 1549 | 172 | 50.93 | 1.75 | 27.14 | 32.3 |
Expected Behavior
latency should increase exponentially
Actual Behavior
latency is saturating
Possible Fix
Steps to Reproduce
- send rate increases from 1-100tps for sample ethereum network
- tx number - 2000
Existing issues
- [ ] Stack Overflow issues
- [ ] GitHub Issues
- [ ] Rocket Chat history
caliper v0.4.2
Context
Your Environment
- Version used:
- Environment name and version (e.g. Chrome 39, node.js 5.4):
- Operating System and version (desktop or mobile):
- Link to your project:
I guess that you got this error because you are using one worker.
The latency is calculated for only successful transactions.
This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.