Performance & limits

Throughput

Throughput measures how much work a system handles over time, such as publications, delivery attempts or bytes processed per second.

Also found under: Operations per second, OPS, Message rate

How it works

Specify the unit before comparing figures. Ten publications with one recipient each create a different workload from ten publications with thousands of recipients. High throughput also does not imply low latency: a system may process substantial work while individual requests wait.

In a Pubb integration

Pubb's operation capacity accounts for publications and delivery attempts, while event bandwidth is another limit. Plan for fan-out and peak publishing bursts as well as your average event rate.

Explore plans and capacity

A practical example

A dashboard publishing five updates per second to one hundred listeners can generate five hundred subscriber delivery attempts per second, in addition to its publication work.