2 years ago

A few days ago, a pending pull-request was merged by mistake. This job had an incomplete implementation and was reverted once the team noticed it was pushed. Now we have a queue that has over 100,000 jobs that resolve a class that doesn't exist, because the PR was reverted. Workers keep picking those jobs up and erroring because they can't resolve the class the jobs belong to.

Now when you deploy your code, workers will pick those jobs up and process them successfully and immediately.