>From GitHub's billing docs, Copilot code review consumes both AI credits
and GitHub Actions minutes (which are free for public repos.)

The AI credit quota depends on ASF's GitHub/Copilot billing setup, which
does not seem to be publicly documented.

On Fri, Jul 17, 2026 at 11:22 AM Kevin Liu <[email protected]> wrote:

> Good to see this is now enabled by ASF! I wanted to give this a try before
> but it wasn't available yet in the Apache org. I think it's very helpful as
> a first review for PRs.
>
> One thing though:
> > before enabling such features, projects should discuss the
> workflow/resource impact with the project team
>
> What's our quota for Copilot Reviews? Will it eat into our CI resources?
> What happens if we use it up?
> Would be good to know these before enabling it. I could find anything from
> the ASF docs above
>
>
> On Thu, Jul 16, 2026 at 7:52 PM vaquar khan <[email protected]> wrote:
>
>> +1
>> Good one 👍
>>
>> Regards,
>> Viquar Khan
>>
>> On Thu, Jul 16, 2026, 8:33 PM Junwang Zhao <[email protected]> wrote:
>>
>>> +1, this can help catch bugs or inconsistencies early in the PR process.
>>>
>>> On Thu, Jul 16, 2026 at 11:20 PM Jones, Danny <[email protected]>
>>> wrote:
>>>
>>>> +1 on trialling something like GH Copilot for review. I was thinking
>>>> about proposing this for iceberg-rust – for the same reasons others raised
>>>> on reviewer bandwidth. I look forward to hearing how it goes for
>>>> iceberg-cpp, and wonder if anyone on the Arrow project can share how it’s
>>>> been for them so far.
>>>>
>>>>
>>>>
>>>> One thing raised by Andrew Lamb on one of the Arrow tuning PRs[1] is to
>>>> define how contributors should respond to the feedback from Copilot. I
>>>> think this would be useful for us too – in particular, I’d encourage
>>>> contributors and reviewers alike to be vocal on the comments in PRs such
>>>> that this can be fed back in for tuning the custom guidance – positive and
>>>> negative! Treat it a bit like providing feedback for a retrospective.
>>>>
>>>>
>>>>
>>>> Danny
>>>>
>>>>
>>>>
>>>> [1]: https://github.com/apache/arrow/pull/50117
>>>>
>>>>
>>>>
>>>> *From: *Renjie Liu <[email protected]>
>>>> *Reply to: *"[email protected]" <[email protected]>
>>>> *Date: *Thursday, 16 July 2026 at 07:04
>>>> *To: *"[email protected]" <[email protected]>
>>>> *Subject: *RE: [EXTERNAL] [DISCUSS] Enable GitHub Copilot code review
>>>>
>>>>
>>>>
>>>> *CAUTION*: This email originated from outside of the organization. Do
>>>> not click links or open attachments unless you can confirm the sender and
>>>> know the content is safe.
>>>>
>>>>
>>>>
>>>> +1 for this. Reviewer bandwidth is limited and if it works well for
>>>> iceberg-cpp, I'm looking forward to introduce it to other repos.
>>>>
>>>>
>>>>
>>>> On Thu, Jul 16, 2026 at 1:18 PM Gang Wu <[email protected]> wrote:
>>>>
>>>> Hi all,
>>>>
>>>> I would like to discuss enabling ASF-managed GitHub Copilot code review
>>>> for
>>>> Iceberg repositories, starting with iceberg-cpp as a small trial:
>>>>
>>>> https://github.com/apache/iceberg-cpp/pull/833
>>>>
>>>> This would enable automatic Copilot review on non-draft PRs, using the
>>>> ASF
>>>> `.asf.yaml` `github.copilot_code_review` setting.
>>>>
>>>> Apache Arrow recently enabled and tuned this feature:
>>>> https://github.com/apache/arrow/pull/50117
>>>>
>>>> ASF Infra documents this as a supported `.asf.yaml` feature:
>>>> https://github.com/apache/infrastructure-asfyaml#copilot-code-review
>>>>
>>>> The `.asf.yaml` docs also say that before enabling such features,
>>>> projects
>>>> should discuss the workflow/resource impact with the project team:
>>>>
>>>> https://github.com/apache/infrastructure-asfyaml#before-you-start-using-asfyaml
>>>>
>>>> I suggest starting with iceberg-cpp first as a trial, because reviewer
>>>> bandwidth
>>>> is limited there and Copilot may provide useful first-pass feedback. If
>>>> it works
>>>> well, other Iceberg repositories can adopt a similar setup later.
>>>>
>>>> Any concerns or objections?
>>>>
>>>> Thanks,
>>>> Gang
>>>>
>>>>
>>>
>>> --
>>> Regards
>>> Junwang Zhao
>>>
>>

Reply via email to