Network Operations Cockpit (Phase 11)

IMPLEMENTED

The Network Operations Cockpit provides ISP NOC engineers, network managers, and support technicians with unified visibility across hardware infrastructure and subscriber sessions without direct device access from browser frontends.

---

1. Architectural Principles

1. Frozen Subsystems (Phase 0–10):
   • Built directly on top of the established MikroTikService, OLTSystemService, ONUService, and NetworkSyncJob service layer.
   • No secondary MikroTik or OLT clients are created.
2. State Decoupling & Caching:
   • PostgreSQL: Authoritative source of truth for persistent device models, subscriber state, POP branches, and durable sync jobs.
   • Redis: Ephemeral high-frequency telemetry caching (cockpit:dash:<tenant>:<pop>:<area>, TTL = 15s) ensuring responsive dashboards under high operator concurrency.
3. Strict Network Isolation:
   • Web frontend clients connect exclusively to Django REST framework endpoints. Frontend code never connects directly to MikroTik RouterOS or OLT Telnet/SNMP ports.
4. Tenant & Area Scoping:
   • Every metric, session, hardware element, and operational job is strictly partitioned by request.tenant and optionally filtered by POP branch or customer area_zone.

---

2. Top-Level Operational KPI Layout

The top of the cockpit renders real-time counts across the network:

• Routers: Total configured MikroTik BNG devices, Healthy (Online), Degraded (Offline or Error status), and aggregate CPU/Memory/Disk meters.
• Customers: Total subscribers, Online (subscribers with active UserSession), Offline, Active, and Expired.
• Pending Actions: Count of in-flight Celery jobs (NetworkSyncJob.JobStatus.PENDING or PROCESSING).
• Failed Actions: Count of failed network synchronization tasks requiring NOC operator review.
• OLT: Healthy and Degraded optical line terminals.
• ONU: Online, Offline, and Optical Alerts (ONUs reporting Rx power < -24.00 dBm, DyingGasp, or Los).
• Live Bandwidth Meter: Concurrent PPPoE sessions, cumulative download/upload bytes, and aggregate bandwidth throughput.

---

3. Hierarchical Operational Workflows

The cockpit is organized into three operational drilldowns:

A. Router → Customers

• Hardware Telemetry: Real-time CPU, RAM, Disk utilization, RouterOS firmware version, effective host, and uptime.
• Provisioned Customers: Table of subscribers provisioned on the selected router with package speeds, expiration dates, and live online/offline badges.
• Active PPPoE Sessions: Live sessions table detailing leased IP addresses, MAC addresses, connection uptime, and byte counters.

B. OLT → ONUs

• Optical Signal Distribution Histogram:
  • Normal (>= -24 dBm): Fiber links within standard specifications.
  • Warning (-27 dBm to -24 dBm): Marginal signal attenuation; requires cable inspection.
  • Critical / LOS (< -27 dBm or DyingGasp / Los): Optical loss or power disruption.
• PON Port Distribution: Active and offline ONU distribution across PON frames (gpon 0/1, epon 0/8, etc.).
• Connected ONUs: Live optical diagnostics detailing serial numbers, MAC addresses, calibrated Rx/Tx power, and distance in meters.

C. Customer → Service Diagnostics

Inspects individual subscriber operational health:

• BNG Session State: Leased IP, MAC, caller ID, session uptime, and active connection timestamp.
• Optical Link: Linked OLT name, PON port, ONU index, optical signal dBm, and distance.
• Sync Job Audit: History of asynchronous sync tasks and error logs.
• Operator Interventions:
  • Disconnect Session: Drops the active session from MikroTik RouterOS and cleans the UserSession record.
  • Sync Speed Profile: Re-pushes bandwidth limits and PPP profiles to the router.
  • Reboot ONU: Dispatches a reboot command to the subscriber's optical network unit via OLT Telnet/SNMP.

---

4. API Endpoints

All endpoints require authentication and strict RBAC (CanViewNetworkMetrics for telemetry, CanControlDevices for control operations):