Cloudflare is well known for DNS, CDN and edge computing, but it is also capable of hosting application infrastructure, particularly API services and processing pipelines.
The advantage of using Cloudflare, as opposed to AWS or GCP, is that more of your plumbing sits in a single environment that is optimised for web performance.
Cloudflare also benefits from a modern DX where services are preconfigured with sensible defaults. For a modern dev team this means less time spent second guessing the cloud configuration and more time spent on design and testing.
Running your API services on Cloudflare workers gives you high-throughput APIs without having to configure machine environments, load balancing, scaling or DNS:
Cloudflare queues give you async processing with automatic retries and a dead letter queue. A worker can be configured to ingest messages from the queue and auto-scale in response to queue length. In my experience, setting up a worker that consumes messages from a queue is very simple compared to AWS (CloudFormation, SQS, IAM, Lambda) and is configured in code using a wrangler file.
AWS and GCP are fine and often necessary - the point of this article is to explain that Cloudflare can handle a lot of application logic and the advantage of doing so is that you get access to Cloudflare's web performance and a modern DX.
George