They corrupt easily and there's no repair method built into journalctl (or anywhere else, really).
The format also does not have any kind of indexing or cache and this is the reason `systemctl status <service>` can take 15s to show the last 10 log lines if you have a few GBs worth of logs.
Well, if you are asking for every log of a given service no matter when was it made, it sure will take time - and I don’t see how indexing or cache would help it (create one specific for each service? That would in practice multiple the size of the logs)
Specify that you only want to last day or so (which is quite likely what you actually want), but I do agree that the default could be something sane like only the last week.
