Add vLLM's health and metrics endpoints

[QwertyJack]

Fix #260

• /health: reuse vLLM's endpoint with a reassigned variable.
• /metrics: reuse vLLM's endpoint using mount_metrics.

[nskumz]

but the metrics were not enabled. everything was #ed. to get these in prometheus these needs to be enabled.

# HELP vllm:num_requests_running Number of requests currently running on GPU.
# TYPE vllm:num_requests_running gauge
# HELP vllm:num_requests_waiting Number of requests waiting to be processed.
# TYPE vllm:num_requests_waiting gauge
# HELP vllm:num_requests_swapped Number of requests swapped to CPU.
# TYPE vllm:num_requests_swapped gauge
# HELP vllm:gpu_cache_usage_perc GPU KV-cache usage. 1 means 100 percent usage.
# TYPE vllm:gpu_cache_usage_perc gauge
# HELP vllm:cpu_cache_usage_perc CPU KV-cache usage. 1 means 100 percent usage.
# TYPE vllm:cpu_cache_usage_perc gauge
# HELP vllm:cpu_prefix_cache_hit_rate CPU prefix cache block hit rate.
# TYPE vllm:cpu_prefix_cache_hit_rate gauge
# HELP vllm:gpu_prefix_cache_hit_rate GPU prefix cache block hit rate.
# TYPE vllm:gpu_prefix_cache_hit_rate gauge
# HELP vllm:num_preemptions_total Cumulative number of preemption from the engine.
# TYPE vllm:num_preemptions_total counter
# HELP vllm:prompt_tokens_total Number of prefill tokens processed.
# TYPE vllm:prompt_tokens_total counter
# HELP vllm:generation_tokens_total Number of generation tokens processed.
# TYPE vllm:generation_tokens_total counter
# HELP vllm:time_to_first_token_seconds Histogram of time to first token in seconds.
# TYPE vllm:time_to_first_token_seconds histogram
# HELP vllm:time_per_output_token_seconds Histogram of time per output token in seconds.
# TYPE vllm:time_per_output_token_seconds histogram
# HELP vllm:e2e_request_latency_seconds Histogram of end to end request latency in seconds.
# TYPE vllm:e2e_request_latency_seconds histogram
# HELP vllm:request_prompt_tokens Number of prefill tokens processed.
# TYPE vllm:request_prompt_tokens histogram
# HELP vllm:request_generation_tokens Number of generation tokens processed.
# TYPE vllm:request_generation_tokens histogram
# HELP vllm:request_params_best_of Histogram of the best_of request parameter.
# TYPE vllm:request_params_best_of histogram
# HELP vllm:request_params_n Histogram of the n request parameter.
# TYPE vllm:request_params_n histogram
# HELP vllm:request_success_total Count of successfully processed requests.
# TYPE vllm:request_success_total counter

[QwertyJack]

If you see nothing during a fresh start, try sending your first request and check again.

[nskumz]

its worked.Thanks