Monitoring ONUs

ONUs are where FTTH problems become subscriber calls, and most of a NOC’s day-to-day in NetSense happens at the ONU level. This guide covers the core workflows: finding an ONU, reading its health, watching it live, and understanding the semantics behind the numbers.

Finding an ONU

ONU search is global: search by serial number or MAC and NetSense finds the ONU across every OLT in the tenant, no need to know which OLT it registered on (useful precisely because subscribers move and splitters get re-patched). Partial serials work, so the number from the subscriber’s router sticker is enough.

Within one OLT, the ONUs tab of the OLT detail view lists everything registered there, with per-ONU columns for weekly/daily availability, Rx/Tx optical power, attenuation, voltage, bias current, temperature, daily QoE, traffic and packet counters, serial/MAC, and Last Down (a relative “x min ago” column that makes recent droppers jump out).

The ONU detail page

An ONU’s page has Overview and Syslog views. Overview carries sub-tabs for Traffic, Packets, QoE and Temperature, a health panel, and a time-range selector; per-ONU history goes back in time, so “was this subscriber’s line already degrading last week?” is answerable.

The Syslog view shows the ONU’s own event stream: register/deregister, dying gasp, port LOS. This is the fastest way to distinguish “power failure at the subscriber” from “fiber problem” on a specific complaint.

Live traffic

The Live Traffic pop-up streams an ONU’s In/Out bits per second in real time, with a selectable refresh of 20, 40 or 60 seconds, the tool for “run your speedtest now, I’m watching” moments with a subscriber on the phone. Use Chrome or Edge for it.

ONU status: how fresh is it?

Real-time ONU up/down depends on syslog from the OLT. With syslog enabled, status changes arrive as they happen; without it, ONU status falls back to the SNMP poll cycle (roughly every 30 minutes), which is the usual explanation when a customer says “the ONU went down but NetSense showed it late”. If that’s your symptom, start with Sending device syslog to NetSense.

ONU online/offline counts refresh about every 3 minutes.

Down reasons, and what “Unknown” means

Where the OLT reports a reason (dying gasp = power loss at the subscriber), NetSense shows it, and power-related reasons are captured reliably. A down reason of “Unknown” means the OLT’s syslog didn’t include one, and in practice an Unknown is almost always fiber-related (cut, macro-bend, splitter issue), especially on EPON OLTs. Treat “Unknown” as “check the fiber path”, not as missing data.

Many ONUs dropping at once with the same timestamp points at a shared cause: a PON port, splitter or feeder cut. The dashboard widget of recent ONU-down counts by reason exists exactly to make that pattern visible at a glance.

Counting and retention semantics

Two deliberate design decisions worth knowing so the numbers don’t surprise you:

• There is no “total ONUs including offline” figure. NetSense counts ONUs it has seen online; it cannot distinguish a permanently removed ONU from one that’s temporarily powered off, so it doesn’t guess.
• Offline ONUs are purged after the retention window: default 7 days, configurable per tenant (some operators run 14). Until then they remain visible as offline. The same 7-day principle applies to down ports/links, which can also be removed manually earlier.

When ONUs are missing or wrong

• Only a fraction of an OLT’s ONUs showing up, while other tools look fine: OLT shows only a fraction of its ONUs.
• ONU list stale or discovery lagging: enable OLT syslog (above), and on VSOL/Syrotech platforms check the Alarm Profile, without it the OLT never emits ONU events at all.
• A device’s Poller tab shows poll interval, duration, queries per cycle and error rate, the direct health readout for that OLT’s SNMP when you suspect polling trouble.