It's good to opt for Lambda in those cases when you expect to have a zillion of computing requirements. If you wish to run some heavy tasks according to the schedule or have an unstable high load during business hours, then Amazon Lambda is the right solution because you pay only for what you use. You will receive the required computing power when it's needed and then your function may 'go sleep' if there is nothing to do.
It's also recommended to use Amazon Lambda solution when you are about to use a lot of services from the same vendor. If you like their architecture, ecosystem, and solution in general, then Amazon Lambda is the choice for you on the way of scalability and flexibility. This method will help you save money and carry out experiments on architecture, technology, and strategy.
*Using Amazon Lambda leads you up to follow all the vendor's rules and meet their limitations.
You may choose technologies only from their stack or pull resources externally.
The development process requires magic and patience sometimes. Lack of debugging and testing tools disappointing. Not all the parts of functionality can be tested locally, so consequently, you're forced to deploy the code and check everything remotely multiple times per day. And that can fill you to the point of agony.
Before you start, we recommend paying attention to the fact that Amazon Lambda as a stack has own limitations:
- Maximum execution duration (per request)
- Memory allocation range
- Ephemeral disk capacity
- Amount of resources
- etc.
Amazon Lambda Limits: https://docs.aws.amazon.com/lambda/latest/dg/limits.html