Durable Functions serve to define complex workflows using code. Anything from monitoring workflows to order processing to batch processing data, durable functions is a viable candidate to consider.
The Application Patterns highlighted in the docs can definitely fit many use cases either alone or in combination with one another.
Surely, the same can be implemented using similar job frameworks run by queues (like celery or bull), but durable functions allow the same to run and scale as a serverless service.
While I'm unaware of any public data of customers using it in production that I can share, the queries raised on durable functions across community forums should be a good indicator of people using it.