We are extensively using multibranch pipelines to build our system and we're struggling to manage the sheer quantity of build logs generated.

On inspection we're seeing roughly 250M of logs (uncompressed) for every build. The compress-build-log-plugin would seem to be the answer and indeed having added this to the properties of the pipeline the log file is compressed to about 2.5M. But, the individual stage logs are not compressed and so are consuming roughly 125M as the logs are entirely replicated so that the pipeline UI can show logs for each step.

Should the compress-build-log-plugin be extended to also find all the *.log files and compress them as well? Not sure what the pipeline UI would think to this though.