With the serverless computing service provided by AWS Lambda, you may run code without setting up or maintaining servers. Manage runtimes easily, maintain event integrations, and provide workload-aware cluster scaling logic. You may execute code for almost any application or backend service with Lambda, all while paying only for the resources you really need. Your fees are determined by the volume of calls made to your functions and the time it takes for your code to run.
When Lambda begins to run in response to an event notification trigger, such as one from Amazon Simple Notification Service (SNS) or Amazon EventBridge, or an invoke call, such as one from Amazon API Gateway, or through the AWS SDK, including test invokes from the AWS Console, a request is counted.
The duration is estimated to be the closest 1 ms* from the start of your code’s execution until it returns or otherwise finishes. The cost is determined by how much memory you give your function. You select the RAM you need for your function under the AWS Lambda resource model, and corresponding CPU power and other resources are allotted. A more considerable memory size makes more CPU available to your function. Consult the Function Configuration manual for further information.
Both x86 and Arm architecture-based processors may be used to run Lambda functions. Compared to functions operating on x86 processors, AWS Lambda functions running on Graviton2, which uses an Arm-based processor architecture created by AWS, provide up to 34% higher pricing performance. The processing of data and media, as well as online and mobile backends, fall under this category of serverless workloads.
Both codes that execute within a function’s handler and initialization codes specified outside the handler are subject to duration charges. Duration for Lambda functions using AWS Lambda Extensions now takes into account the time needed for the shutdown phase of the last running extension’s code to complete. Consult the documentation for the Lambda Programming Model for further information.
One million free requests per month and 400,000 GB-seconds of computing time per month are included in the AWS Lambda free tier and may be used for functions running on both x86 and Graviton2 processors together. Lambda additionally provides tiered price alternatives for on-demand duration choices beyond specific monthly consumption limits.
AWS Lambda takes part in Compute Savings Plans, a flexible pricing structure that provides reduced costs on utilization of Amazon Elastic Compute Cloud (Amazon EC2), AWS Fargate, and Lambda in return for a commitment to a steady quantity of consumption (calculated in $/hour) over a one- or three-year period. You may save up to 17% on AWS Lambda with Compute Savings Plans. Duration and Provisioned Concurrency are both subject to savings.