Documentation is the map
You can't operate what you can't see. Network docs are topology diagrams, an IP address plan (IPAM), naming standards, port maps, rack elevations and standard configs. When a diagram disagrees with reality, the diagram is a liability - keep it current or troubleshooting starts from a lie.
Monitoring and baselines
Watch latency, packet loss, jitter, bandwidth, interface errors and CPU/memory. The trick is the baseline: 'normal' from last month is what makes today's spike obvious. Alerts without baselines cause alert fatigue - everything looks urgent, so nothing is. NetFlow shows who is talking to whom and how much.
Change management
Risky changes go through a process: a maintenance window, an approval, pre-checks (capture the working state), the implementation, post-checks (prove it still works), and a written rollback plan in case it doesn't. The goal is to make change boring and reversible, not heroic.
Backups, lifecycle and disaster recovery
Archive device configs so a failed box can be rebuilt fast. Track firmware/software versions and end-of-life/end-of-support dates so you're not running gear that can't get security fixes. DR planning uses two numbers: RPO (how much data you can afford to lose - drives backup frequency) and RTO (how long you can be down - drives your recovery design).