Request

Because of some jvm bug, running async-profiler sometimes can cause jvm crash async-profiler/async-profiler#1100 (comment)

A workaround is using --cstack dwarf or --cstack vm to the argument to avoid calling AsyncGetCallTrace.

And actually since async-profiler4.0, it is recommended to use --cstack vm. async-profiler/async-profiler#1168 (comment)

Can we add support for passing --cstack argument in to pyroscope.java component? Either adding it to profiling config or hard coding it is fine.

Use case

after running alloy in production 4 days, 1 application pod (30 in total) crashed because of the issue. it would be nice if alloy can support --cstack configuration together with bumping async-profiler version.