Serverless architectures are special with their stateless, event-driven, and distributed nature. For this reason, developers should put special attention to monitoring them throughout their software development cycle not after they face a problem in production.
However, implementation of monitoring and observability on serverless is a lot more complex. As a developer, there are some best practices that you can follow while implementing your own instrumentation but all of them pollutes your business logic. Therefore, it is wise to use an already existing solution like Thundra to gather valuable insights out of your serverless architecture.
This whitepaper provides:
An overview of serverless use cases which require in-depth monitoring.
Differences between monitoring and observability and what those mean in serverless.
Tips on implementation of observability yourself with manual and/or automated instrumentation.
Best practices of serverless observability such as asynchronous monitoring, sampling, etc.