SNMP Poll
SNMP Poll periodically reads scalar OIDs or walks one subtree through an SNMP Device and emits value-change and availability events.
Poll the smallest useful data set
Create an SNMP Device and confirm its read operations first. Add SNMP Poll beneath the device, choose scalar or tree mode, and start with a conservative interval.
Choose scalar mode for known values
Use a comma- or newline-separated list when the required OIDs are already known. One multi-GET is usually cheaper and more predictable than walking a broad tree.
Bound tree walks
Use tree mode for a single subtree and set a realistic maximum row count. Avoid walking a device root on every cycle.
Set failure thresholds
Use consecutive failure and recovery thresholds to avoid availability flapping during isolated packet loss.
Configuration parameters
| Parameter | Required | Description | Default |
|---|---|---|---|
Title | Yes | Display name for the poll plan. | None |
SNMP device | Yes | Active source SNMP Device. | Parent device |
Mode | Yes | SCALAR uses multi-GET; TREE uses GETBULK walk. | SCALAR |
OIDs | Yes | Comma- or newline-separated scalar OIDs, or one subtree root in tree mode. | None |
Interval | Yes | Poll interval from 1 to 86400 seconds. | 30 seconds |
Maximum rows | Tree | Maximum rows returned by one subtree walk, from 1 to 10000. | 200 |
Publish unchanged | Yes | Emits a value event on every successful cycle instead of only after a change. | Disabled |
Failure threshold | Yes | Consecutive failures before the task becomes unavailable. | 3 |
Recovery threshold | Yes | Consecutive successes before an unavailable task returns to healthy. | 1 |
Separate values from availability
SnmpValueChangedEvent
Contains OID, name, type, old value and new value. The first normal poll establishes a baseline.
SnmpPollStateEvent
Reports UNAVAILABLE after the failure threshold and HEALTHY after recovery.
Protocol health
Availability is based on the SNMP result rather than ICMP reachability.
The first failed cycle does not immediately emit an availability transition. This prevents a single startup timeout from becoming an alarm before the configured threshold can be evaluated.
Control load and event volume
Multiply interval frequency by device count and OID count when estimating load. Broad walks, short intervals and Publish unchanged can generate substantial network traffic and event volume. Use Event Manager filters and history retention deliberately.