Size SNMP polling load: varbinds per second, PDUs per cycle, bandwidth and daily data points.
Varbind rate is the total object count divided by the interval — the sustained work the poller must do. GETBULK packs many varbinds into one response, so PDU count depends on max-repetitions, and bandwidth doubles because every request has a matching reply. Poller capacity is measured in varbinds per second, and shortening a five-minute interval to one minute multiplies load fivefold on both the collector and every polled agent's CPU.
SNMP Polling
poll load = OIDs × devices ÷ poll interval, giving varbinds retrieved per second.
GETBULK request count
PDUs per cycle = devices × ceil(OIDs ÷ max-repetitions)
poll load = OIDs × devices ÷ poll interval, giving varbinds retrieved per second. Varbind rate is the total object count divided by the interval — the sustained work the poller must do. GETBULK packs many varbinds into one response, so PDU count depends on max-repetitions, and bandwidth doubles because every request has a matching reply.
Poller capacity is measured in varbinds per second, and shortening a five-minute interval to one minute multiplies load fivefold on both the collector and every polled agent's CPU.
This calculator takes 5 inputs: Monitored devices, OIDs polled per device, Poll interval, Bytes per varbind, GETBULK max-repetitions. The pre-filled defaults are a realistic starting point — replace them with figures from your own environment for a result you can act on.
Each PDU costs a round trip and an agent CPU cycle. Raising max-repetitions cuts the number of exchanges dramatically, but set it too high and responses exceed the path MTU and fragment or get dropped — 25 to 50 is the usual safe range.