Operational Monitoring and Maintenance
Watching the device
Core operational commands: show version, show chassis alarms, show chassis routing-engine, show system uptime, show interfaces terse (one line per interface with admin and link state), show interfaces extensive (errors, drops, queue counters), show arp, show route and show route forwarding-table (the FIB as the PFE sees it), show log messages, show log chassisd. Pipe filters sharpen output: show interfaces | match up, | except, | count, | no-more.
Real-time tools
monitor interface ge-0/0/0 gives a refreshing rate display. monitor traffic interface ge-0/0/0 gives a tcpdump-style capture (add no-resolve to skip DNS lookups) and you must name the interface. monitor start messages tails the log live. ping and traceroute are available as source-address-aware operational commands, and clear interfaces statistics resets counters.
Maintenance tasks
request system snapshot copies / and /config to /altroot and /altconfig, producing a recovery image. request system storage cleanup frees space; show system storage reports usage. request system reboot and request system power-off do the obvious. request system zeroize wipes configuration and files and returns the device to factory state - destructive. request system configuration rescue save stores a known-good rescue config that you can restore with rollback rescue. file list, file copy and file show manage local files, and request support information collects troubleshooting output.
Traps
- Confusing snapshot (non-destructive backup) with zeroize (destructive reset).
- Assuming interface counters are cumulative forever; they reset on clear or reboot.
- Thinking monitor traffic captures all interfaces by default.
Rule of thumb
When something is broken, work outward: show chassis alarms, show interfaces terse, show log messages, then drill into the specific subsystem.