while ago we had discussed the issue that you can't get error output from an ongoing workflow with failed jobs. At the moment this is what our expert users are doing to get around the problem that pegasus-analyzer is rarely useful for our workflows (almost all problems are identified in large / long running workflow where it will never show useful output when we want to debug failed jobs). However, I think manually going to the log files is not ideal for new users.
Is there any progress on a solution to this, or should we develop one of our own? The only thing I think I could do is put together some scripts to search and dissect the log files.