---
type: Article
title: "Monitor a UPS via SNMP — UPS-MIB Setup"
description: "A UPS with a network card answers the standard UPS-MIB (RFC 1628, 1.3.6.1.2.1.33): read battery status, runtime left and load over SNMP on any brand."
resource: "https://snmp-monitoring.com/devices/ups/"
tags: [devices, snmp]
timestamp: 2026-07-13T00:00:00Z
---

# Monitor UPS via SNMP

An uninterruptible power supply earns its keep in the seconds after the mains drops — and those are exactly the seconds you need to know about. A proper SNMP UPS monitoring device tells you it's now running on battery, how many minutes of runtime are left, and whether the battery itself is healthy enough to deliver them. The good news is that there's an actual standard for this. Any UPS with a network card that implements the standard UPS-MIB answers the same core OIDs, so you're not stuck reverse-engineering a proprietary format. This guide covers what SNMP support looks like, which values to poll, how to enable it, and why an external watcher is the whole point during a power event.

## SNMP support on UPS

Most UPS units gain SNMP through a **UPS SNMP card** — a Network Management Card (NMC) that slots into the unit and gives it an Ethernet port and its own IP address. Smaller consumer units often skip the card and instead connect by USB to a host running software like Network UPS Tools (NUT), which can then re-expose the readings to a monitoring system. Either way, the endgame is the same: a pollable agent that reports the power picture.

There's a genuine standard here, which is rare and welcome. **RFC 1628 defines the UPS-MIB**, rooted at `1.3.6.1.2.1.33`. A UPS whose card implements it answers a consistent set of objects for battery, input, output, and alarms regardless of brand. In practice most vendors implement the standard UPS-MIB *and* layer a richer enterprise MIB alongside it, so the RFC 1628 objects are your portable, safe baseline and the vendor MIB is where the extras live.

On versions, UPS network cards commonly support **SNMPv1** and **SNMPv2c**, and newer cards add **SNMPv3**. As with any device, a v2c community travels in clear text, so keep the card on a management network and reach for v3 where the card supports it and the path isn't trusted. Note that this page is the device guide for the UPS itself — for the general power/battery *sensor* readings and thresholds, the [UPS sensor page](/sensors/environmental/ups.md) owns that topic.

![Diagram of an external SNMP poller reading UPS-MIB battery, input, output, and alarm groups from a UPS SNMP card over UDP 161](https://snmp-monitoring.com/img/devices/ups-snmp-poll.webp)

## What you can monitor on UPS

The standard UPS-MIB (RFC 1628, subtree `1.3.6.1.2.1.33`) organizes readings into named groups. Rather than fabricate column numbers, the honest and portable way to describe it is by those groups — poll them by name once you've loaded the UPS-MIB into your tool:

| Group (under `1.3.6.1.2.1.33`) | What it reports |
|---|---|
| Battery | On-battery vs on-mains status, estimated minutes of runtime remaining, charge remaining, battery voltage/temperature |
| Input | Incoming mains voltage and frequency per input line |
| Output | Output voltage, current, and **load as a percentage** per output line |
| Alarm | Active well-known alarms — on battery, low battery, replace battery, and similar conditions |

The three readings you'll alert on most:

- **Power source / on-battery status** — the single most urgent flag. The instant it flips to battery, a power event is underway.
- **Estimated runtime remaining** — how many minutes the battery can hold the load. This is your countdown to a graceful shutdown.
- **Output load percentage** — a UPS overloaded past its rating will cut runtime dramatically and may trip; watch it as a capacity-planning number too.

Battery temperature and per-cell health are also exposed and worth trending, since heat is what quietly kills a battery long before it fails outright — the [temperature sensor page](/sensors/environmental/temperature.md) covers thermal monitoring in general. A battery that tests fine today but runs a few degrees hotter each month is telling you something the "on battery / on mains" flag never will, and it's the kind of slow drift that only shows up when you keep the history. Anything beyond the RFC 1628 objects — deep per-string diagnostics, self-test history, model-specific counters — lives in the **vendor's enterprise MIB**, and those numeric OIDs vary by manufacturer and firmware. Load the vendor MIB and walk your own card rather than trusting a copied value.

## Enabling SNMP — steps

Setup happens on the UPS network card, through its own web interface, not on a server. The steps are consistent across cards:

1. **Give the card network access.** Assign it an IP (static is best for a device you'll poll), and connect its management port to your management VLAN.
2. **Log into the card's web UI** with its admin credentials and open the SNMP settings.
3. **Enable the SNMP agent** and pick the version — SNMPv2c for a trusted management network, SNMPv3 if the card supports it and the path is exposed.
4. **Set a read-only community string** — long and random, never `public` — and, where the card allows, restrict polling to your collector's IP.
5. **Confirm UDP 161** is permitted to the card and not blocked by an intervening firewall.
6. **Verify from a workstation** by walking the UPS-MIB subtree:
   ```bash
   snmpwalk -v2c -c <community> 10.0.0.20 1.3.6.1.2.1.33
   ```
   You should see the battery, input, output, and alarm objects come back. To read numerically when the UPS-MIB isn't installed locally:
   ```bash
   snmpwalk -v2c -c <community> -On 10.0.0.20 1.3.6.1.2.1.33
   ```

Keep the card firmware current — UPS management cards have had their share of security fixes — and treat the community like the password it effectively is. The broader [setup section](/setup/index.md) covers hardening and the collector side.

## Monitoring it externally

Here's the uncomfortable irony of powering your monitoring stack from the same UPS it's supposed to watch. If a prolonged outage drains the battery, the UPS, the server, *and* the collector all go down together — and the alert you needed most, "runtime critically low, shutting down," is the one that never gets sent. A watcher that shares the failure can't report it.

This is precisely the [double durability](/monitoring/double-durability.md) argument, and it's sharpest for power. [ostr.io Monitoring](https://ostr.io/service/monitoring) polls your UPS card's SNMP agent from outside your network — zero-setup SaaS, nothing to install on the card — and sends real-time email and SMS alerts when it flips to battery, when runtime drops toward zero, or when it stops answering entirely. Because those checks run on independent infrastructure over a separate path, they keep working while your site is dark. Point ostr.io at your card's public [monitoring](/monitoring/index.md) endpoint and you get warning you can act on, delivered to a phone that isn't plugged into the same wall.

![Chart of UPS estimated runtime minutes and output load percentage during a mains power event](https://snmp-monitoring.com/img/devices/ups-runtime-chart.webp)


## FAQ

**Do I need a special card to monitor a UPS over SNMP?**
Usually yes — a Network Management Card (the UPS SNMP card) adds the Ethernet port and agent. Smaller units without a card slot can be connected by USB to a host running software like NUT, which then exposes the readings to your monitoring system.

**Is there a standard MIB for UPS devices?**
Yes. RFC 1628 defines the UPS-MIB, rooted at `1.3.6.1.2.1.33`, with standard groups for battery, input, output, and alarms. A card that implements it answers the same core objects across brands; vendor MIBs add extras on top.

**Which UPS readings should trigger an alert?**
Three above all: on-battery status (a power event just started), estimated runtime remaining (your shutdown countdown), and output load percentage (overload risk). Battery temperature is worth trending because heat shortens battery life.

**Can I poll the UPS-MIB without vendor MIBs installed?**
Yes. Walk `1.3.6.1.2.1.33` and use the `-On` flag for numeric output if the UPS-MIB isn't installed on your machine. You only need the vendor MIB for manufacturer-specific objects beyond the RFC 1628 baseline.

## Conclusion

Monitoring an SNMP UPS monitoring device is more standardized than most hardware: enable SNMP on the UPS network card, set a scoped read-only community, and poll the RFC 1628 UPS-MIB at `1.3.6.1.2.1.33` for battery status, estimated runtime, output load, and alarms. Lean on the vendor MIB only for extras, and on SNMPv3 where the path isn't trusted. Most of all, poll the card from outside — because a UPS running down its battery is the last thing that can afford a monitor sharing its power.