In order to truly embrace the pragmatic [[Serverless-first mindset]], you need to have an awareness of use cases where serverless is not a good fit. This question is a moving target as workload use cases that in the past were in the "unsuitable" bucket are being gradually removed as cloud providers create new services and expand the capabilities of their existing ones. ## List of workloads unsuitable for public cloud serverless - Where a workload cannot be run in the public cloud (e.g. it needs access to on-premise firewalled systems) - Where a single long-running process is required. In these cases, consider [[AWS Fargate]]. [[AWS Lambda|Lambda]]'s current max duration is 15 minutes. - Certain specialised databases such as [[Neo4j]] and [[Elasticsearch]] do not currently have a fully serverless offering. While there are hosted managed services available for these, you still need to worry about managing your own scaling and you pay a fixed hourly rate based on your peak load. - Workloads that require consistent low-latency performance, such as high-frequency trading applications in banks or submillisecond robotics automation in warehouses, are not good candidates for event-driven architecture. [^fn1] [^fn1]: [AWS Lambda Operator Guide](https://docs.aws.amazon.com/lambda/latest/operatorguide/intro.html) by [[@James Beswick]] ([[AWS Lambda]]) --- tags: #Serverless