Add new finish reasons introduced in TensorRT-LLM v0.16.0.

The short arg of `prometheus_port` conflicts with `port`. Remove theshort arg variant.Fixeshuggingface#3205

When a request is cancelled, the `tensorrt_llm::executor::Result`contains `outputTokenIds` with size 1, but `outputTokenIds[0]` has size0. This causes `as_generation_step` to segfault.Check the size of `outputTokenIds` and `logProbs` before attempting toaccess the inner vector. The `finishReasons` can be skipped because ithas only one dimension and the minimum beam size is 1.Because cxx have not added Option support yet, include two boolean flagsto denote whether the value is valid.Change log level when request is cancelled to debug.

Support per request stop sequences.

The trycatch only uses the `what()` method, which means we can catch thebroader `std::exception` instead. This is beneficial becausenlohmann/json also throws exception.

When the required config files are not present, nlohmann/json throwsparsing error, which does not help much for identifying what was wrong.Check the existence of these files early and return specific errormessages.

Currently, the do_sample option is ignored and the executor will alwayssample. Set top_k to 1 if do_sample is false.

Get a more accurate inference start time from the trtllm response.Because `Instant` does not expose absolute value, create referencepoints on both sides and return duration relative to the referencepoint instead.

The executor_status_looper runs a spin loop, even if there are no activerequests. This makes the service constantly wasting a CPU core.Make the loop block on receiving requests if there are no running onesto reduce CPU usage when idle.