Ask HN: What is your LLM monitoring stack?

Consider using resources like TensorFlow Profiler, PyTorch Profiler, and Spark MLlib's profiler. you can also take advantage from the capabilities of tracing libraries like TensorBoard or Prometheus.