TSV and you're done smaller readable, esp. with `column -t <log` works with awk/cut/join/grep/sort/column/etc./etc./etc. if you have complicated enough logs

TSV and you're done smaller readable, esp. with `column -t <log` works with awk/cut/join/grep/sort/column/etc./etc./etc. if you have complicated enough logs that you can't maintain the shell scripts that parse them, you probably also have enough log data that json's going to blow up your space and you probably want indexes anyway, so throw it in a real database (oh hi I work for one of these, log analysis is actually one of our strong suits) but others have already commented to this effect