Linux Command Reference ยท Systemd and Logging (Read-Only)
Log priorities (severity levels)
The eight syslog severity levels from emerg (0) to debug (7), how journalctl -p, logger -p and rsyslog selectors use them, and which level to choose when logging.
Every syslog and journal message carries a priority, a number from 0 to 7 that says how serious it is. The scale is shared by the kernel, syslog, logger, journalctl, rsyslog and libraries such as syslog(3). Lower numbers are more severe, which surprises many people: priority 0 is a catastrophe and 7 is debugging chatter.
When you filter by priority you select a threshold: journalctl -p warning shows warning (4) and everything more severe (3, 2, 1, 0). You can run the real commands only against a live journal, so those are read-only here, but the examples work with the numbers and with logger's level names.
journalctl -p err -b logger -p user.warning "disk 90% full" dmesg --level=err,warn
Try it
Press Run (or Ctrl + Enter) to see the result here.
Program input ยท what you'd type while the program runs
Type the answers before you press Run, one per line. Each time the program reads a line (with read or cat), it gets the next line from this box.
Press Run (or Ctrl + Enter) to see the result here.
Program input ยท what you'd type while the program runs
Type the answers before you press Run, one per line. Each time the program reads a line (with read or cat), it gets the next line from this box.
Syntax
journalctl -p LEVEL # LEVEL or LEVEL..LEVEL; name or number logger -p FACILITY.LEVEL "message" dmesg --level=LIST # kernel messages printk: <3>kernel message # the kernel's own prefix notation
| Number | Name | Typical use |
|---|---|---|
| 0 | emerg |
The system is unusable; kernel panic. Broadcast to all terminals by many configurations. |
| 1 | alert |
A human must act at once, for example the main database is corrupt. |
| 2 | crit |
Critical failures: hardware errors, a service's core function lost. |
| 3 | err |
Errors: an operation failed but the service continues. |
| 4 | warning |
Something unexpected that may become a problem. |
| 5 | notice |
Normal but noteworthy events: a service started, a configuration reloaded. |
| 6 | info |
Routine progress information. |
| 7 | debug |
Detail useful only when developing or diagnosing. |
Returns: journalctl -p err is shorthand for -p 3; the accepted aliases are emerg, alert, crit, err, warning, notice, info, debug (and the old synonyms error, warn, panic in logger).
Errors and edge cases: the numeric value is not the same as the syslog PRI header, which combines facility and level as facility * 8 + level (for example auth.err = 4 * 8 + 3 = 35).
Usage notes
- Lower is worse.
-p 3includes 0, 1 and 2. To see only exactly one level in journalctl use a range such as-p 4..4. - Pick the level by who must react, not by how you feel:
errmeans somebody should look soon,warningmeans watch it,noticeandinfoare for the record. - Default thresholds hide debug. Most services log
infoand above in production;debugis switched on per service, often with a flag orSYSLOG_LEVEL. - Kernel log level (
/proc/sys/kernel/printk) controls which kernel messages reach the console;dmesgstill shows all of them. See The /proc and /sys filesystems. - Do not log secrets at any level, and avoid
debugoutput in production services that handle personal data. - systemd services can set
SyslogLevel=andLogLevelMax=; lines printed with a<N>prefix on stdout (for example<4>cache low) are logged at that priority. - Alerting rules usually trigger on
error worse; counting them withjournalctl -p err --since today | wc -lis a quick health check.