Syslog
When a router’s interface flaps or a switch’s CPU spikes at 2:00 a.m., the evidence is written as a syslog message — if you configured logging. Syslog is the standard way network devices record events and push them to a central collector, and understanding its severity levels and message format is the difference between reading a log line and decrypting it. This lesson covers the 0–7 severity scale, message anatomy, IOS logging destinations, and how to verify the whole pipeline.
- Recall the eight syslog severity levels (0 emergency … 7 debugging) and what triggers each.
- Describe the syslog message format: facility, severity, mnemonic, and description.
- Configure the IOS logging destinations: logging host, logging trap, and logging buffered.
- Verify logging with show logging.
- Read a real log line and explain what a timestamp, severity, and mnemonic tell you.
Severity levels 0–7
Every syslog message carries a severity from 0 (most severe) to 7 (least). The exam is merciless here: you must know all eight, in order, including their names and numbers.
| Level | Keyword | Meaning | Example |
|---|---|---|---|
| 0 | emergency | System unusable. | Router hardware failure, no usable memory. |
| 1 | alert | Immediate action needed. | Temperature alarm, power loss imminent. |
| 2 | critical | Serious condition. | Interface loss that isolates a site. |
| 3 | error | Error condition. | Failed authentication, null TTL dropped. |
| 4 | warning | Warn of future trouble. | High CPU or memory utilisation. |
| 5 | notification | Normal but notable event. | Configuration applied from console. |
| 6 | informational | Normal informative event. | Interface up, adjacency formed. |
| 7 | debugging | Verbose debugging output. | debug ip packet output. |
logging trap 4 means
level 4 and more severe (0–4) are forwarded; a message at level 5 is
below the threshold and is not forwarded. “Higher number, lower urgency” is the
mental model.
The severity also behaves hierarchically: configuring a destination for severity 4 captures warnings, errors, critical, alert, and emergency messages because their numbers are ≤ 4. Debugging (7) is deliberately the noisiest and should never be shipped to production collectors.
Anatomy of a syslog message
A Cisco syslog message has a fixed, structured head. Take this line:
*Jun 11 14:22:31.408: %SYS-5-CONFIG_I: Configured from console by console
Reading it left to right:
- Timestamp —
*Jun 11 14:22:31.408. The leading asterisk means the clock is not synchronised with an NTP source; when NTP (Lesson 10) locks, the asterisk changes to a date and time form that matches the NTP reference. Accurate timestamps depend entirely on the time discipline you configured in Lesson 10. - Facility —
SYS. The subsystem that generated the message (SYS = system, IF = interfaces/line protocols, LNK = link, ETH, CDP, OSPF, SNMP, and so on). - Severity —
5. Here a notification. - Mnemonic —
CONFIG_I. A short machine-readable name for the event. CONFIG_I means someone applied a configuration change (the “I” distinguishes it from the longer CONFIG_IP logs of old IOS). - Description —
Configured from console. The human-readable explanation that follows the colon.
Facility-Severity-Mnemonic is the address of an event in the Cisco logging
taxonomy: which subsystem (SYS), how urgent (5), and which named event
(CONFIG_I). Recognising this triplet lets you parse any log line even when the
description is unfamiliar.
A classic event to recognise is the interface flap pair —
%LINK-3-UPDOWN: Interface GigabitEthernet0/1, changed state to down followed by
%LINEPROTO-5-UPDOWN. Those two mnemonics (UPDOWN appearing in both LINK and LINEPROTO
facilities) are a running joke in the exam because they appear in almost every infrastructure scenario.
Where logs can go
IOS has three logging locations worth knowing (a fourth, the console, exists but is covered in the console-logging basics). Each has a different value:
- Console (default) — messages print to the console line. Useful for a directly-connected setup, useless when nobody is looking, and it can interrupt interactive work.
- Buffer (in RAM) — the router keeps messages in an internal, circular RAM
buffer. Nothing leaves the device; you inspect it with
show logging. Survives until reboot only. - Syslog server (remote) — messages are pushed over the network to a central collector, typically on UDP 514 (and TCP/TLS variants). This is the enterprise answer: one place sees every device, and the logs survive the losing device.
- SNMP — logging events can be mirrored as SNMP traps (Lesson 11) into the NMS, tying the two management protocols together.
logging host in enterprise designs.
Configuring logging on IOS
A minimal, realistic remote-logging configuration:
R1(config)# logging host 192.168.1.100 R1(config)# logging trap informational R1(config)# logging source-interface gigabitEthernet 0/0 R1(config)# logging facility local7
Decode each line:
logging host 192.168.1.100— the collector’s address. Repeat the command for redundant collectors.logging trap informational— forward severity 6 and more severe (0–6) to the remote host. The keyword names the threshold; remember the direction (6 and more severe) from Section 1.logging source-interface— fixes the source address of log packets, the same trick you used for NTP. Useful across NAT and access lists.logging facility local7— tags the messages with the facility value carried in the UDP datagram (local0–local7 are reserved for user processes). Collectors use it to sort or filter.
If you also want the on-box buffer with its own threshold (so show logging shows
something even when the collector is unreachable):
R1(config)# logging buffered 65536
logging buffered is the buffer size in bytes
(here 65,536) — an exam distractor will offer “65536 messages” or forget the size
entirely. If you want a severity on the buffer, add it: logging buffered 65536 informational.
And a running debug with a low threshold will fill any buffer in seconds — keep
debugging out of the forwarded thresholds.
Verifying with show logging
show logging is the one-stop verification command: it reports the configuration, the counters, and the buffered messages:
R1# show logging Syslog logging: enabled (11 messages dropped, 4 messages rate-limited, 0 flushes, 0 overruns, xml disabled, filtering disabled) Logging to: 192.168.1.100 (udp port 514, audit disabled, link up, 3 messages logged, 0 lines logged, 0 xml rejected) Log Buffer (65536 bytes): *Jun 11 14:21:19.192: %LINK-3-UPDOWN: Interface GigabitEthernet0/1, changed state to up *Jun 11 14:22:31.408: %SYS-5-CONFIG_I: Configured from console by console
show logging reveals
The first line answers “is logging even enabled?” The Logging to: block
proves the collector line is up and shows how many messages made it. The buffer shows actual
messages with timestamps, facility-severity-mnemonic, and description — exactly the reading
skill from Section 2. If a scenario shows a collector that never receives logs, the three suspects
are: logging disabled, threshold filtering the events, or UDP 514 blocked between the devices.
The drop and rate-limit counters in the first line are quietly important: IOS rate-limits chatty messages and drops when the pipeline can’t keep up. Non-zero drop counters with a healthy-looking config usually mean a burst of debugging or a badly-chosen threshold.
Reading logs in practice
The skill the exam actually tests is decoding a line. Walk through three real shapes:
%LINK-3-UPDOWN: Interface GigabitEthernet0/1, changed state to down %LINEPROTO-5-UPDOWN: Line protocol on Interface GigabitEthernet0/1, changed state to up %SYS-5-CONFIG_I: Configured from console by vty0 (192.168.1.50)
- The first says physical layer: interface G0/1 went administratively or physically down (severity 3, error).
- The second says line protocol: after the physical link recovered, the data-link came up (severity 5, notification). One event typically produces both, in that order.
- The third is a configuration change — and notice it also records who and from
where (
vty0 (192.168.1.50)). In an incident, that is the audit trail of who touched the box. Combined with a matching NTP-synchronised timestamp it is exactly the kind of evidence a security investigation reconstructs.
* before the timestamp means the device clock is not NTP-synchronised —
the timestamps are only as trustworthy as the internal clock. If you see * on logs you
are supposed to correlate, the fix is not log parsing, it is Lesson 10’s
ntp server command. Synchronised devices print the date and time with no asterisk, and
your forensic timeline suddenly lines up.
Troubleshooting decks love the summary: severity 0–2 need immediate attention, 3 is an actual
error, 4 is the run-up to trouble, 5–6 are the normal rhythm of the network, and 7 is
debugging noise. When an interface flaps and you must prove whether it is physical or the protocol,
read the facility and mnemonic: LINK-3-UPDOWN versus LINEPROTO-5-UPDOWN is
the definitive answer.
The vocabulary you need
Every term below appears in the lesson text above. Select one to read its definition, then use the quiz at the end to check what stuck.
Review
Check what you have learned with these review questions.
Sources & references
- Cisco. CCNA (200-301) Exam Topics. cisco.com.
- IETF. Request for Comments (RFCs). rfc-editor.org.
- IANA. Internet Assigned Numbers Authority registries. iana.org.
Target exams: CCNA 200-301 v1.1 · CyberOps Associate 200-201. Last content verification: 2026-10-11.