More documentation on signature and threads management

[pierre-yves-monnet]

Describe the bug
More documentation, please

Why does the Handler method have an async? Is that mandatory? At high throughput, it leads to multiple problems. Do you think we can remove the async?
2/ Customer creates a .NET application with 60 workers. From time to time, it seems when a worker finishes a batch of 15, it waits 5 seconds before getting another batch, even if there are a lot of tasks.
When in the configuration, we specify 15 threads, is this 15 threads PER WORKER or for ALL THE 60 workers?

To Reproduce
No bug, more documentation

Expected behavior
A section in the documentation to explain if

Environment (please complete the following information):

Additional context

[ChrisKujawa]

Hey maybe next time we can use the Discussions for such https://github.com/camunda-community-hub/zeebe-client-csharp/discussions

Sounds to me more like some questions/discussion topics and not an actual issue to fix.

1. Why does the Handler method have an async? Is that mandatory? At high throughput, it leads to multiple problems. Do you think we can remove the async?

The job handler supports both (async and sync) to avoid such discussions and make the implementation easy as possible.
https://github.com/camunda-community-hub/zeebe-client-csharp/blob/main/Client/Api/Worker/IJobWorkerBuilderStep1.cs#L55-L104

Async/await is a common pattern in the .net world. If you see any weird behavior I doubt it is because of this, likely it is something different.

2/ Customer creates a .NET application with 60 workers. From time to time, it seems when a worker finishes a batch of 15, it waits 5 seconds before getting another batch, even if there are a lot of tasks.
When in the configuration, we specify 15 threads, is this 15 threads PER WORKER or for ALL THE 60 workers?

Not sure I can completely follow. Would be great if can share the complete configuration. But it is not unexpected that it might take a while until it polls the next tasks, depends on the configuration, and it might be even that tasks already locked by other workers. To completely answer this I need more context and details.