With _LOGGER_DISCOVERY_DONE=true, log_to_journal will never attempt rediscovery again in the same shell session, even if logger becomes available later (installed, PATH fixed, etc.) or if LOGGER_PATH was blanked due to an earlier transient failure. If that 'no recovery until re-source' behavior is not intended, consider a retry strategy (e.g., allow periodic re-discovery when LOGGER_PATH is empty, or reset _LOGGER_DISCOVERY_DONE when LOGGER_PATH is empty and a call is at/above CURRENT_LOG_LEVEL).
Originally posted by @Copilot in #90 (comment)