Played with #Knative on #Kubernetes today, it's pretty nifty
I used `oha` to stress-test a basic health-check service, and it's cool that #KnativeServing can scale from 0 service pods, to 1 pod, and eventually up to 10 pods, all without dropping any requests (the requests are queued if there aren't enough service pods)
It's like having my very own AWS Lambda at home (and honestly easier for me to use than AWS Lambda)
These are the results for running `oha -c 1000 -z 30sec` against my cluster which has 3x Raspberry Pi 4B nodes (all nodes are control-plane nodes but with pod-scheduling enabled):
```
Summary:
Success rate: 1.0000
Total: 30.0055 secs
Slowest: 6.2945 secs
Fastest: 0.0264 secs
Average: 0.5445 secs
Requests/sec: 1793.4723
Total data: 0 B
Size/request: 0 B
Size/sec: 0 B
Response time histogram:
0.146 [2198] |■■■■
0.292 [10205] |■■■■■■■■■■■■■■■■■■■■■■
0.438 [14373] |■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■
0.584 [9585] |■■■■■■■■■■■■■■■■■■■■■
0.731 [10099] |■■■■■■■■■■■■■■■■■■■■■■
0.877 [4081] |■■■■■■■■■
1.023 [1561] |■■■
1.169 [495] |■
1.315 [257] |
1.461 [140] |
1.607 [820] |■
```