[Discussion] Optimize Doc Bot

[Yu <li...@apache.org>]

Hi Pulsarers,

As you know, to improve the efficiency of documentation management, we
introduced the doc Bot to our community last year [1]. It has been run for
more than 8 months.

During this period, we’ve found some potential and serious issues in doc
Bot, which lowers productivity. Consequently, we initiate this discussion
to optimize it.
# Why optimize doc Bot?

Issues in doc Bot:


   -

   Hard to maintain and nearly impossible to scale.

Doc Bot was inspired by auto-labeling [2] and implemented by just modifying
the compiled dist/index.js rather than src, which means if we want to add
or modify some features (eg. forced doc label checks), it’s nearly
impossible.


   -

   Doc Bot can not function correctly in certain scenarios.

Bug details are recorded here [3].
# How to optimize doc Bot?

After thinking twice, we plan to implement a new doc Bot in Golang to
replace the current one (doc-label-check) [4] in GitHub Action.

Feel free to check the prototype and try it here [5].

~~~~~~~~

Please review and comment within 72 hours. If there is no discussion or
objection, we’ll implement it as planned, thank you!

[1]
https://docs.google.com/document/d/1bQfZkSu5nG1tNycpmXXtUFn-Z5-h-uqHv6IXsCEySQ8/edit


[2] https://github.com/harupy/auto-labeling

[3] https://github.com/apache/pulsar/issues/15797

[4] https://github.com/apache/pulsar-test-infra/tree/master/doc-label-check

[5] https://github.com/maxsxu/action-labeler

Yu and Max