read-journal
Reads from the systemd journal and emits entries as events. A cursor and timestamp are stored in db so that processing resumes from where it left off.
When tail-interval is set, the plugin re-reads the journal on that interval until shutdown. Without it, the plugin reads once (backfill) then exits.
Config
Optional:
journalctl: Path to the journalctl binary (default:journalctl)db: State file path (default:logbus-journal-db)since: Start time as Unix epoch seconds (default: now). Ignored once a cursor is stored indb.batch-size: Maximum journal entries per invocation (default:1000)tail-interval: Seconds between re-reads (default: none — read once and exit)
Example
pipeline:
read-journal:
config:
db: /var/lib/logbus/journal-state.json
since: 1700000000
tail-interval: 60