Use danger standalone? (danger triggers github rate limit exceeding)

[divramod]

Hey,

i am running danger standalone in a kubernets pod-container.
The command i use to run danger looks like that:

export DANGER_GITHUB_API_TOKEN=${GIT_TOKEN} 
danger pr https://github.com/${GIT_OWNER}/${GIT_REPO}/pull/${GIT_PULL_REQUEST}

This is running fine and gives the expected results.
My problem is, that our repo seems to be really large and while running that process danger makes about 1900 calls to the github api, which lets us run into the rate limit really fast.

Now i wanted to ask, if i can somehow avoid that?
I am not really sure where to start here to avoid the github api calls.

thx in advance!

[orta]

This sounds like the code in the dangerfile is doing some overly intense work - honestly, I'd start there - is it accidentally reading every file in the repo somehow?

[divramod]

hey @orta, thx for the reply.

yeah, the dangerfile seems to do a lot of comparing-work.
but it seems, it should be that way.

can you think of another solution for tackling this problem.

i could think of downloading the branch/commit which danger compares to and do the heavy work locally?

is there a option going into this direction?

i am open to every solution which avoids penetrating the github api.

[orta]

Try wrap up #991 which verifies that you're running on the right commit, and uses the fs instead of the GitHub API for read access to files - that only works if you're running danger on CI, but that's at the 90% use case