Network Optimization
A scheduled, read-only audit of your network gear that produces a best-practice report with a score and prioritized findings. It's strictly informational — it never changes anything, opens tickets, or sends alerts. It exists to answer: "how healthy and well-configured is my network, really?"
Admin-only: the tab appears on the Dashboard only for admin users.
What it scans
Network gear only — gateway/router, switches, and access points, plus the features they handle (VLANs, SSIDs, WAN). Endpoints, servers, and client devices are not scanned.
- UniFi gear — device health, port/link status, WAN, VLANs, SSIDs (from the controller)
- SNMP devices — interface status, errors/discards, speed/duplex (works for other vendors too)
- Reachability/ports — a light
nmappass (-T3, capped, no intrusive probes)
The appliance itself is always excluded (self-protection) — it can never scan itself.
Running a scan
- Run now — the button on the Network Optimization tab. A scan of a typical home network takes 1–3 minutes; you can watch progress and cancel.
- Scheduled — enable a recurring schedule (day + hour, local time) or a one-time run. Default suggestion: weekly at 03:00.
The score
The report scores your network 0–100 (baseline 100, penalties per finding):
| Severity | Impact | Example |
|---|---|---|
| Critical | −20 each (big drop) | A device unreachable that should be up |
| Warning | −5 each (stacks) | HTTP management exposed, repeated link flaps |
| Info | capped (small, never tanks the score) | Single uplink, unnamed ports, unused VLANs |
A healthy home network should score 90+ — a low score means real config concerns to look at, not noise. Performance and security are scored separately so you can see where the points went.
Findings
Findings are grouped by severity and category:
- Performance — link speed/duplex mismatches, interface errors, congestion, high CPU/mem
- Security — exposed SSH/HTTP management, weak/legacy settings, stale firmware
- Reliability — link flaps, single WAN/uplink, persistent issues
- Hygiene — unnamed ports, unused VLANs, disabled SSIDs, stale records
Each finding has a stable key, a severity, and the evidence behind it — so you can act on it (and later releases can link findings to fixes).
Safety guarantees
- Read-only — nothing is changed on your gear
- No alerts — findings never open tickets or send emails
- Self-protection — the appliance excludes itself, always
- Conservative scans — capped hosts, light timing, no intrusive probes; you control the scan profile, concurrency, and schedule