How does AWS Lambda charge for compute time?

AWS Lambda charges for compute time based on the duration that your code is actively running in response to an event or request. This means that you are billed solely for the time your function executes, measured in milliseconds, from the moment it begins until it stops executing. This model is beneficial because it allows you to efficiently manage costs; you do not incur charges when your code is not running, thus emphasizing the event-driven nature of AWS Lambda.

The other options do not accurately reflect AWS Lambda's pricing model. Charges based on server uptime would imply ongoing fees even when the code isn't actively executed, which is not applicable since Lambda operates serverlessly and scales automatically. A flat monthly fee would not align with usage-pricing models based on specific metrics like execution time. Finally, while AWS Lambda does charge per request in terms of invocation count, compute pricing is specifically tied to the duration of code execution, not merely the act of receiving requests.