hongchaodeng

he3.app founder here :) Thanks for sharing it! I couldn't believe someone shared the tool I built on a website that I browsed a lot. I'm the founder and also doing the marketing/development. I have dreamt about posting it some date on HN once it is more powerful and popular, but was too shy and busy to do it at the moment. I'm really happy to see someone else share it which I should have done earlier and maybe more. Thanks again! That really gives me courage to continue the work.

Hi, he3.app founder here. Thanks for using it. We built He3 to provide long-tail tools -- tools that might be helpful to developers but no big-enough market to have a business to support it. Yes I'm a Chinese. But I used to work in the silicon valley and we built it initially in China for labor cost reasons. Eventually we plan to benefit developers around the world and would need people/community outside China to achieve that :D

Exactly on our roadmap :D

- Linux: coming soon. It is built using Electron. We started with Mac and just finished testing on Windows. We will make our CI to test on Linux and publish a version for Linux soon.

- Open Source: We have plans for open sourcing it. But the code is not ready yet. Will prepare for open sourcing once it gets more stable.

Unleashing rate limiters is like releasing a water gate -- now you know which part is too weak to stop the flood.

By doing this, we can have more insights of system performance and find out which part is too "weak".

There is no reason that we couldn't increase the rate limiter if we keep improving the system in overall.

Thanks for reading it so carefully. Let me explain it one by one:

> they optimized "Round" which took 18 seconds

This is CPU profile. The overall sampled time is 79 seconds. The Round() method took 18 seconds. This has no direct indication to scheduling latency.

> by that reduced total cpu time from 53 to 23 seconds

This is about scheduling latency. In this experiment, we schedule 1000 pods on 1000 nodes and see how much time it took to schedule a pod, aka scheduling latency. Before the optimization, the latency is 53 seconds; after, 23 seconds. Of course, this is just one of the steps we kept optimizing.