Situation:
- Serverless functions run on vendor-managed back-end environments so that IT can focus on building business logic, not provisioning and deploying servers.
- High availability and capacity management is left to the vendor, and auto-scaling means handling bursting demand is yesterday’s problem.
- With precise per-second billing, you truly pay only for the resources you use.
Complication:
- Higher degrees of abstraction in cloud services require changing skill sets for operations staff. It becomes difficult to distinguish what tasks remain important, which ones can be eliminated, and what skills need to be added.
- Stateless applications, such as AWS Lambda, instantiate and disappear after running each request. In this disparate environment, it hard to know not only how to manage but what to manage.
- Making the case for operational staff becomes critical when the business falls into the NoOps trap created by vendors.
Resolution:
- By effectively monitoring and securing your serverless environment, the operations team enables developers to write valuable functions for the business and help mitigate the exposure to risk.
- While technology may change, the need for resiliency and disaster recovery does not. You must build resiliency for your functions and for the data that functions require when they are invoked.
- Many infrastructure metrics may be resolved but cost optimization is more important than ever. In fact, by observing those costs, you may realize that it makes more sense to build out serverless functionality on your own private cloud.