[Feature] add CUDA support in the metrics service

[LorenzoTettamanti]

Is your feature request related to a problem?

• Yes

  Problem Description:

  As the documentation and the README announced, the GPU support will be introduced in the 0.1.6 version. GPU's are an essential part of Machine learning and AI so a support to trace CUDA calls must be introduced

  Solution:

• Introduce a GPU support using uprobes and uretprobes directly in the libcudart.

• Trace essential functions calls like cudaFree,cudaMalloc,cudaMemcpy

• Add new OTEL metrics (must be discussed together)

• Adapt the existing consumer library to consume the GPU extracted data correctly in the pipeline

  Additional info

  common/src/otel-metrics.rs: Contains all the OpenTelemetry metrics instrument to process the data from the eBPF perf buffers
  common/src/semantics.rs: Contains all the semantic conventions for the metrics service
  common/src/consumer.rs: Consumer library to read to consume the data that comes from the perf-buffers

---