ULTIMATE SNMP
POLLER
SNMP was supposed to be replaced "soon" after 1988. That was 37 years ago. USP bridges the gap between this improbably resilient protocol and modern observability platforms — without the existential dread.
You Know This Pain
SNMP has been obsolete since approximately the moment it was invented. Yet it remains the primary management protocol for network devices everywhere. Modern observability platforms? They treat SNMP like an afterthought.
Proprietary MIB Formats
Your vendor requires custom YAML profiles. Standard ASN.1 MIBs don't work. New device? Submit a ticket and wait.
One Interval Fits All
You need interface counters every 30 seconds, but entity inventory every 5 minutes. Too bad — it's one interval per device.
Collector Islands
Each poller is configured independently via config files. No central management. No visibility. Pure chaos at scale.
Single Backend Lock-in
Your collector only speaks to one platform. Want to send data to both Elastic and Datadog? Run two collectors.
Actually Works™
Not "works most of the time" or "works if you configure seventeen things correctly." It genuinely, improbably, works.
Adaptive Intelligence
Legacy pollers hammer devices until they break. USP listens. The agent calculates EWMA of response times, dynamically adjusting polling frequency.
- Auto-tuning intervals per device
- Fast lanes and slow lanes for traffic
- Host health tracking isolates problems
Unbreakable Resilience
Network partitions happen. Backends fail. Your monitoring shouldn't stop.
- Circuit breakers detect backend failures
- Durable disk-backed buffering
- Rendezvous hashing for automatic failover
- No single point of failure
Standard MIB Support
Upload any ASN.1 MIB file. It compiles. It works. Even the custom one you wrote 12 years ago.
- Full SNMPv1, v2c, v3 support
- One-click MIB import and compilation
- Automatic dependency resolution
- O(1) OID lookups
Zero-Friction Deployment
No JVM. No Python venv. No external databases. One static Go binary for the Controller. One for the Agent.
- Embedded React WebUI
- Single binary per component
- ~20MB memory footprint
- Deploy in minutes, not days
Multi-Backend Output
Send data to one platform or all of them simultaneously. Your observability stack, your choice.
- Elasticsearch & OpenSearch
- Datadog, New Relic, Splunk
- Prometheus & OpenTelemetry
- Simultaneous multi-destination
OIDyssey™ Explorer
Stop guessing OIDs. Walk live device trees through the UI. Smart classification recommends metric types automatically.
- Live SNMP walks from the browser
- Auto-detect Counter vs Gauge
- One-click template creation
- Full MIB tree visualization
How USP Stacks Up
Side-by-side with the tools you're probably already frustrated with.
| Capability | USP | DataDog | Splunk | ElastiFlow |
|---|---|---|---|---|
| Standard ASN.1 MIBs | Profiles | YAML | ||
| Per-MIB Polling Intervals | ||||
| Central Controller | ||||
| Multi-Backend Output | Simultaneous | Native only | Native only | ES only |
| Adaptive Polling | EWMA-based | Partial | ||
| Agent HA / Failover | HRW Hashing | |||
| WebUI Management | Full GUI | YAML files | ||
| External Dependencies | None | Agent required | Heavy Forwarder | Multiple |
See It In Action
A real interface for real network engineers. Not a config file with aspirations.
Dashboard
Fleet health, output pipelines, agent status at a glance
OIDyssey Walker
Live SNMP walks with automatic MIB translation
Host Management
Per-host integration assignments with independent intervals
MIB Explorer
Browse compiled MIBs, search OIDs, create templates
Works With Your Stack
Native connectors. No middleware. No custom scripting.
Configure multiple backends per agent. Send interface metrics to Datadog for real-time alerting and to Elasticsearch for long-term analysis. Simultaneously. Because that's how real infrastructure works.
In Which Someone Finally Gets Annoyed Enough
We were running an enterprise SNMP solution that cost more than some of our engineers. The engineers were more useful, but the software had better contract terms.
It used proprietary YAML profiles instead of standard MIBs. It couldn't poll IF-MIB at 60 seconds and ENTITY-MIB at 5 minutes on the same device. Each collector was an island — not in the tropical paradise sense.
USP was built to solve a real problem: get SNMP data into modern observability platforms without the pain, the vendor lock-in, or the six-figure price tag.
"There is a theory which states that if anyone ever discovers the perfect SNMP integration, it will instantly disappear and be replaced by something even more inconceivable. Another theory states this has already happened several times."— Perpetual Obsolescence Technologies
Ready to Fix Your SNMP Problem?
Free version includes 100 hosts, 5 custom collections, and 1 tenant. No time limit. No credit card. One zip file. Five minutes to your first metrics.
Questions? Concerns? Existential dread about SNMP? → support@perpetual-obsolescence.tech