On what basis is AWS Lambda's billing calculated?

AWS Lambda's billing is calculated based on two main factors: the number of requests your function handles and the duration of code execution. Specifically, AWS Lambda charges you for each request that your function receives and for the compute time that your code takes to execute.

The number of requests is simply the count of how many times your function is called. Meanwhile, the duration is measured in milliseconds, starting from the moment your code begins execution until it finishes or returns a response. This model allows for a highly scalable and flexible pricing structure, where you only pay for the resources you actually use, rather than for reserved capacity or a flat monthly fee.

By focusing on usage, AWS Lambda supports a pay-as-you-go model that is beneficial for applications with variable usage patterns. This means you avoid the costs associated with pre-provisioned infrastructure while only incurring costs directly correlated to your application's activity.