worqcloud
Capacity planning

Serverless Concurrency Calculator

Estimate average and burst concurrency from request rate and execution time, then add planning headroom.

Your workload

Results update as you edit. The starting values are illustrative.

Arrival rate for one function or route.
Time each invocation remains in flight, including dependency waits.
Peak arrival rate divided by average rate. Assumes duration stays the same.
Extra capacity above the modeled burst.
Your current or proposed capacity, rather than an assumed provider quota.
YOUR ESTIMATE

90

Concurrent slots for the modeled burst, including headroom.

Average in flight
25
Modeled burst in flight
75
Compared capacity
100
Additional slots for this plan
0

Average concurrency = requests/second × duration/1,000. Multiply by burst and headroom, then round up once for the plan.

How to use this estimate

This model assumes one invocation per request and unchanged average duration during a burst. It estimates steady demand and a chosen peak scenario. It does not simulate queueing, retries, cold starts, provider scaling speed, or a latency guarantee.

Measure actual concurrency and dependency limits before setting capacity. Reserved concurrency and provisioned concurrency have different purposes. Use this result as a load-test starting point.

AWS Lambda concurrency guidance ↗

Read the practical guide →