prio:

backlog:

logger? slog?
error handling
sort by column header not only by data key?
transformer in config -> manipulate source
better in memory persistence? https://github.com/kelindar/column
grok parser?