24/7 MikroTik & OLT Monitoring
The always-on nervous system of my ISP: automated monitoring across 2 locations for 200+ PPPoE clients — Telegram alerts the moment something drifts, GPON OLT diagnostics, QoS classification, and scheduled self-healing automation.
// Overview
Running an ISP as an independent operator means being the NOC engineer, the field technician, and the on-call staff at once. This project is the automation layer that makes that sustainable: a self-hosted Fedora server polling routers and OLTs around the clock, pushing alerts to Telegram and WhatsApp, and executing scheduled maintenance on its own — bandwidth & FUP management, scheduled PPPoE restarts, and daily configuration backups.
The whole infrastructure is reachable remotely through Cloudflare Tunnel and Tailscale, with systemd services, cron scheduling, and automatic restart keeping the stack alive 24/7 without human babysitting.
// The Challenge
A small ISP cannot afford a dedicated NOC, but its customers still expect problems to be noticed before they complain. Outages and optical degradation happen at night; PPPoE congestion builds silently; a stuck customer session means support calls at midnight. Every minute between "something broke" and "someone noticed" is churn risk — so detection had to be automatic, continuous, and mobile.
// What I Built
A layered monitoring and automation stack on a single always-on server:
Router & Client Monitoring
MikroTik API polling across 2 locations — PPPoE session state, bandwidth, and FUP status for 200+ clients.
Telegram & WhatsApp Alerts
Instant alerts with severity context — incidents surface on the phone in my pocket, not on a screen nobody watches.
GPON OLT Diagnostics
Optical signal monitoring and diagnostics for the FTTH plant — degrading ONU light levels flagged before customers lose service.
QoS Classification
FastTrack-based queue trees and traffic classification for fair bandwidth sharing during peak hours.
Scheduled Self-Healing
Scheduled PPPoE restarts, daily router backups, and auto-restarting services — routine maintenance runs without me.
Secure Remote Access
Cloudflare Tunnel + Tailscale for management access from anywhere — no exposed ports, no public management panels.
// Results
The ISP operates with one person and a stack of automations where others need a shift schedule: incidents are detected and alerted automatically, routine maintenance runs on cron, and configuration history survives every failure via daily backups.
2
ISP locations monitored
200+
PPPoE clients tracked
24/7
Automated alerting & backup
0
Exposed management ports
// Tech Stack
Want your network watching itself?
I build monitoring and automation for small ISPs and offices — detection, alerting, and self-healing included.
Get in Touch