We have a pipeline with code along the lines of

tee('output.log') {
  sh 'terraform plan'
}

and we've noticed that output.log is occasionally truncated (the end of the file has a predictable pattern, so we alert on when it's missing), while stdout, as shown in the Jenkins console log, is not truncated.

We've seen this happen even when output.log was not a large file (just a few kilobytes). It seems to happen the most when there's a lot of build congestion (many jobs running concurrently). It doesn't happen often and we haven't found a way to reproduce the issue at will.