Network Management, Automation, and Monitoring
This domain tests how you get configuration onto devices and telemetry off them: CLI and API access, automation tooling, configuration lifecycle, zero-touch provisioning, SNMP, syslog, NTP and flow sampling.
Access and automation
Console and SSH for interactive work; RESTCONF/REST API and NETCONF for programmatic configuration on OS10; Ansible and Python for repeatable rollout. Store intent in version control and push idempotently rather than editing devices by hand.
Configuration lifecycle
Distinguish running-config from startup-config: an unsaved change disappears on reload. OS10 supports candidate configuration with commit and rollback, so a bad change can be reverted without a reboot. Firmware installs typically go through ONIE on open networking platforms; verify the image and the boot choice before reloading.
Zero-touch provisioning
ZTP lets a factory-fresh switch obtain an image and configuration automatically: DHCP supplies the location of the image and a configuration script, and the switch applies them at first boot. Validate with the ZTP status output and by confirming the running-config matches the intended template.
Monitoring
SNMPv2c is simple but clear-text; SNMPv3 adds authentication and privacy. Syslog severities run 0 (emergency, most severe) to 7 (debug); send to a central collector with consistent timestamps from NTP. sFlow samples traffic for flow visibility; streaming telemetry pushes counters continuously. Dell OpenManage Enterprise and SupportAssist cover fleet inventory, firmware baselines and case creation.
Worked example
A new S5248F-ON must be deployed without a console visit. DHCP scope option hands out the image URL and a configuration script; on boot the switch detects ZTP, downloads the OS10 image, installs it, applies the script that sets hostname, management IP, NTP, syslog and AAA, then reboots into the final image. Verification is the ZTP status, the running-config, and a syslog message at the collector.
Traps
Forgetting to save the running-config; reversing syslog severity (0 is most severe, not least); deploying SNMPv3 without auth/priv settings; skipping NTP and then misreading log timelines; hard-coding per-device config instead of templating.