Update Go Modules Runs Causes Rate Limit Issues

[dmikusa]

Expected Behavior

Each repo should have a job to update go modules. It should run successfully.

Current Behavior

It will run for most of our repos, but there is a growing number of repos and what happens is that about 10-20% of the repos will fail when the job runs. This is because all of the jobs are configured to run on a cron schedule which kicks off at exactly the same time. This pushes us over Github rate limits.

Possible Solution

Two potential solutions come to mind:

1. We could use different user tokens. I believe the rate limits are per-user token and all of the jobs run with the same token so as the same user. We could try to break that up and use multiple tokens to spread out the load.
2. We could stagger the jobs by a small amount. Rather than running them all at 2:00 AM Monday, we could stagger that and have them run sometime randomly (or maybe not randomly, maybe some sort of even distribution) between 2:00AM and 2:30AM.

Steps to Reproduce

https://github.com/paketo-buildpacks/dynatrace/actions/runs/3870244851/attempts/1

Motivations

It is a pain to go and re-run all the failed jobs each week. If that doesn't happen, then updates do not get applied to all of the buildpacks.