Should we increase the number of `batch`?

[evshary]

I found that the CI of up-transport-zenoh-rust often fails due to the HTTP 429, which is Too Many Requests.
https://github.com/eclipse-uprotocol/up-transport-zenoh-rust/actions/runs/16766243510/job/47471747243
I guess this might be related to the batch number here.

ci-cd/.github/actions/run-eclipse-dash/check-3rd-party-licenses.sh

Line 33 in 38cfcdb

args=(-jar "$dash_jar" -timeout 60 -batch 90 -summary "$dash_summary")

Do you think we should increase the number here? @sophokles73

[sophokles73]

I can increase the number in my next PR. However, if what you believe is true, then I guess that it will have little impact as the overall number of requests as perceived by Clearly Defined will probably be originating from GitHub altogether, i.e. I do not believe that the request limit is applied at the project level ...

[evshary]

I guess you're right. Here is an issue about this. eclipse-dash/dash-licenses#448
Perhaps we can run the check daily (or even weekly) instead of every PR?

[sophokles73]

We currently run the job when pushing to the main branch. We also use a PR trigger that only kicks in for changes to the Cargo.* files. Maybe we can simply remove the (unconditional) push to main trigger in a first step?

As for running the job on a weekly basis only: my main concern would be to make sure that we are in good shape when creating a release. Maybe we can run weekly (regularly) but also as part of the release process? Which might lead to failed release job runs due to 429 errors as well, though ...

[evshary]

Maybe we can simply remove the (unconditional) push to main trigger in a first step?

Agree with you. Will do that.

Maybe we can run weekly (regularly) but also as part of the release process? Which might lead to failed release job runs due to 429 errors as well, though ...

It will really bother us if the release flow fails due to the check...
If we check the Cargo.* in every PR, I suppose it should be enough to avoid the issue.

[PLeVasseur]

Hey @evshary, @sophokles73 -- checking in here on how we'd like to proceed 🫡

[evshary]

I think we can simply check it only in PR as the first stage. At least removing push to main trigger makes it less likely to happen on up-transport-zenoh-rust. If other projects also have a similar issue, we can consider increasing the batch to see if that helps.