Serverless seems charming both for pre-revenue companies and enterprise-level giant organizations with its pay-as-you-go model. Modern application development practices encourage the adoption of distributed microservices which only works when needed. This leads us to the granular level of cost reduction because we don’t pay to cloud vendors for whole architecture while we generally use only a subset of it. Sounds like a good story right? Pay-as-you-go can cause you a shocking bill if you don’t follow the best practices and keep your eye on the costs. It is worth to say that the guilty part is most of the time not functions themselves but the managed services or third-party APIs or your own logging practices.
Join us as we walk through:
The unique cost problems of serverless.
Best practices to control retry and timeout mechanisms that are unique to serverless.
How to detect the slowdowns in third-party APIs and take precautions.
How to best tune your logging mechanism without they become the biggest cost item in your bill.
How observability can help you save money even when there is no failure at your system.