Calculate partition count, throughput units, and consumer group layout.
Last verified: May 2026
Output will appear here...The calculator takes inputs (peak ingress MB/sec, average message size, expected consumer count, processing latency requirements) and computes the minimum partition count using Event Hub's 1 MB/sec per partition rule, the minimum TU count using the 1 MB/sec ingress / 2 MB/sec egress per TU rules, and the recommended target (with 30% headroom). It surfaces the SKU comparison (Standard vs Premium vs Dedicated) for the recommended capacity.
The Azure Event Hub Partition Calculator helps you determine the optimal number of partitions and throughput units for your Azure Event Hubs deployment. Partition count directly affects parallelism and throughput capacity, and it cannot be changed after creation. This tool calculates recommendations based on your expected message rate, message size, consumer count, and processing requirements, helping you make an informed decision before provisioning.
Your team is provisioning a new Event Hub for IoT telemetry from 10,000 devices, each sending a 500-byte message every 5 seconds. The calculator: peak rate = 10K × 500B / 5s = 1 MB/sec ingress. Recommendation: 4 partitions (more than the minimum 1, room for consumer parallelism), 2 TUs (1 needed + 1 buffer), Standard tier. Total cost: ~$22/month. The team would have over-provisioned to 32 partitions and 10 TUs without the calculator — paying $80/month for capacity they'd never use.
Partition count is permanent on Standard SKU and cannot be changed after creation — but Premium SKU now supports scaling partitions up (not down). If you're not 100% sure of your future scale, start on Premium with 4 partitions and scale up as needed. The premium price is often justified by avoiding a costly migration later.
Always over-provision partitions, not throughput units. TUs can be added/removed dynamically per minute. Partitions are forever. A common pattern: provision 32 partitions even if you only need 4 today, then start with 1 TU and scale TUs as load grows.
Each partition can support up to 1 MB/sec ingress regardless of TU count — partitions are the parallelism unit, TUs are the bandwidth unit. If you need 10 MB/sec total but with 32 consumers reading in parallel, you need at least 32 partitions even if 10 TUs would cover the bandwidth.
Partitions determine the physical data layout and consumer parallelism. Changing the count would require redistributing all data and breaking consumer offset tracking. For this reason, Azure locks partition count at creation time (though Premium/Dedicated tiers now support scaling up). Choose carefully using this calculator.
Each throughput unit (TU) provides 1 MB/s ingress and 2 MB/s egress. Partitions determine the number of parallel consumers. You need enough TUs for total throughput and enough partitions for consumer parallelism. A common starting point is to match partition count to expected peak consumer instances.
Was this tool helpful?
Disclaimer: This tool runs entirely in your browser. No data is sent to our servers. Always verify outputs before using them in production. AWS, Azure, and GCP are trademarks of their respective owners.