When executing a workflow build that is triggered by SCM polling, if a triggered run takes longer than the polling period to complete, then additional runs will be scheduled every time the polling period expires until the initially triggered build completes.

I've only seen this issue with the mercurial plugin so far, but I don't really have any builds that use any other SCM plugins, so that doesn't necessarily mean anything.

A minimal workflow that should recreate the problem is below (fill in your own SCM details):

  //Do an SCM checkout here.

  sh 'sleep 120' // Set the polling interval to 1 minute