# HELP python_gc_objects_collected_total Objects collected during gc
# TYPE python_gc_objects_collected_total counter
python_gc_objects_collected_total{generation="0"} 110037.0
python_gc_objects_collected_total{generation="1"} 20956.0
python_gc_objects_collected_total{generation="2"} 359.0
# HELP python_gc_objects_uncollectable_total Uncollectable objects found during GC
# TYPE python_gc_objects_uncollectable_total counter
python_gc_objects_uncollectable_total{generation="0"} 0.0
python_gc_objects_uncollectable_total{generation="1"} 0.0
python_gc_objects_uncollectable_total{generation="2"} 0.0
# HELP python_gc_collections_total Number of times this generation was collected
# TYPE python_gc_collections_total counter
python_gc_collections_total{generation="0"} 3141.0
python_gc_collections_total{generation="1"} 283.0
python_gc_collections_total{generation="2"} 15.0
# HELP python_info Python platform information
# TYPE python_info gauge
python_info{implementation="CPython",major="3",minor="10",patchlevel="22",version="3.10.22"} 1.0
# HELP process_virtual_memory_bytes Virtual memory size in bytes.
# TYPE process_virtual_memory_bytes gauge
process_virtual_memory_bytes 2.945540096e+10
# HELP process_resident_memory_bytes Resident memory size in bytes.
# TYPE process_resident_memory_bytes gauge
process_resident_memory_bytes 3.688574976e+09
# HELP process_start_time_seconds Start time of the process since unix epoch in seconds.
# TYPE process_start_time_seconds gauge
process_start_time_seconds 1.79205227175e+09
# HELP process_cpu_seconds_total Total user and system CPU time spent in seconds.
# TYPE process_cpu_seconds_total counter
process_cpu_seconds_total 260.01
# HELP process_open_fds Number of open file descriptors.
# TYPE process_open_fds gauge
process_open_fds 38.0
# HELP process_max_fds Maximum number of open file descriptors.
# TYPE process_max_fds gauge
process_max_fds 10240.0
# HELP vllm:num_requests_running Number of requests currently running on GPU.
# TYPE vllm:num_requests_running gauge
vllm:num_requests_running{model_name="study-model"} 0.0
# HELP vllm:num_requests_waiting Number of requests waiting to be processed.
# TYPE vllm:num_requests_waiting gauge
vllm:num_requests_waiting{model_name="study-model"} 0.0
# HELP vllm:num_requests_swapped Number of requests swapped to CPU.
# TYPE vllm:num_requests_swapped gauge
vllm:num_requests_swapped{model_name="study-model"} 0.0
# HELP vllm:gpu_cache_usage_perc GPU KV-cache usage. 1 means 100 percent usage.
# TYPE vllm:gpu_cache_usage_perc gauge
vllm:gpu_cache_usage_perc{model_name="study-model"} 0.0
# HELP vllm:cpu_cache_usage_perc CPU KV-cache usage. 1 means 100 percent usage.
# TYPE vllm:cpu_cache_usage_perc gauge
vllm:cpu_cache_usage_perc{model_name="study-model"} 0.0
# HELP vllm:cpu_prefix_cache_hit_rate CPU prefix cache block hit rate.
# TYPE vllm:cpu_prefix_cache_hit_rate gauge
vllm:cpu_prefix_cache_hit_rate{model_name="study-model"} -1.0
# HELP vllm:gpu_prefix_cache_hit_rate GPU prefix cache block hit rate.
# TYPE vllm:gpu_prefix_cache_hit_rate gauge
vllm:gpu_prefix_cache_hit_rate{model_name="study-model"} -1.0
# HELP vllm:num_preemptions_total Cumulative number of preemption from the engine.
# TYPE vllm:num_preemptions_total counter
vllm:num_preemptions_total{model_name="study-model"} 0.0
# HELP vllm:prompt_tokens_total Number of prefill tokens processed.
# TYPE vllm:prompt_tokens_total counter
vllm:prompt_tokens_total{model_name="study-model"} 43392.0
# HELP vllm:generation_tokens_total Number of generation tokens processed.
# TYPE vllm:generation_tokens_total counter
vllm:generation_tokens_total{model_name="study-model"} 12480.0
# HELP vllm:time_to_first_token_seconds Histogram of time to first token in seconds.
# TYPE vllm:time_to_first_token_seconds histogram
vllm:time_to_first_token_seconds_bucket{le="0.001",model_name="study-model"} 0.0
vllm:time_to_first_token_seconds_bucket{le="0.005",model_name="study-model"} 0.0
vllm:time_to_first_token_seconds_bucket{le="0.01",model_name="study-model"} 0.0
vllm:time_to_first_token_seconds_bucket{le="0.02",model_name="study-model"} 0.0
vllm:time_to_first_token_seconds_bucket{le="0.04",model_name="study-model"} 0.0
vllm:time_to_first_token_seconds_bucket{le="0.06",model_name="study-model"} 0.0
vllm:time_to_first_token_seconds_bucket{le="0.08",model_name="study-model"} 56.0
vllm:time_to_first_token_seconds_bucket{le="0.1",model_name="study-model"} 58.0
vllm:time_to_first_token_seconds_bucket{le="0.25",model_name="study-model"} 90.0
vllm:time_to_first_token_seconds_bucket{le="0.5",model_name="study-model"} 163.0
vllm:time_to_first_token_seconds_bucket{le="0.75",model_name="study-model"} 175.0
vllm:time_to_first_token_seconds_bucket{le="1.0",model_name="study-model"} 192.0
vllm:time_to_first_token_seconds_bucket{le="2.5",model_name="study-model"} 195.0
vllm:time_to_first_token_seconds_bucket{le="5.0",model_name="study-model"} 195.0
vllm:time_to_first_token_seconds_bucket{le="7.5",model_name="study-model"} 195.0
vllm:time_to_first_token_seconds_bucket{le="10.0",model_name="study-model"} 195.0
vllm:time_to_first_token_seconds_bucket{le="+Inf",model_name="study-model"} 195.0
vllm:time_to_first_token_seconds_count{model_name="study-model"} 195.0
vllm:time_to_first_token_seconds_sum{model_name="study-model"} 64.45989656448364
# HELP vllm:time_per_output_token_seconds Histogram of time per output token in seconds.
# TYPE vllm:time_per_output_token_seconds histogram
vllm:time_per_output_token_seconds_bucket{le="0.01",model_name="study-model"} 0.0
vllm:time_per_output_token_seconds_bucket{le="0.025",model_name="study-model"} 0.0
vllm:time_per_output_token_seconds_bucket{le="0.05",model_name="study-model"} 11245.0
vllm:time_per_output_token_seconds_bucket{le="0.075",model_name="study-model"} 12214.0
vllm:time_per_output_token_seconds_bucket{le="0.1",model_name="study-model"} 12214.0
vllm:time_per_output_token_seconds_bucket{le="0.15",model_name="study-model"} 12214.0
vllm:time_per_output_token_seconds_bucket{le="0.2",model_name="study-model"} 12224.0
vllm:time_per_output_token_seconds_bucket{le="0.3",model_name="study-model"} 12240.0
vllm:time_per_output_token_seconds_bucket{le="0.4",model_name="study-model"} 12247.0
vllm:time_per_output_token_seconds_bucket{le="0.5",model_name="study-model"} 12266.0
vllm:time_per_output_token_seconds_bucket{le="0.75",model_name="study-model"} 12269.0
vllm:time_per_output_token_seconds_bucket{le="1.0",model_name="study-model"} 12270.0
vllm:time_per_output_token_seconds_bucket{le="2.5",model_name="study-model"} 12279.0
vllm:time_per_output_token_seconds_bucket{le="+Inf",model_name="study-model"} 12285.0
vllm:time_per_output_token_seconds_count{model_name="study-model"} 12285.0
vllm:time_per_output_token_seconds_sum{model_name="study-model"} 584.5129952430725
# HELP vllm:e2e_request_latency_seconds Histogram of end to end request latency in seconds.
# TYPE vllm:e2e_request_latency_seconds histogram
vllm:e2e_request_latency_seconds_bucket{le="1.0",model_name="study-model"} 0.0
vllm:e2e_request_latency_seconds_bucket{le="2.5",model_name="study-model"} 48.0
vllm:e2e_request_latency_seconds_bucket{le="5.0",model_name="study-model"} 183.0
vllm:e2e_request_latency_seconds_bucket{le="10.0",model_name="study-model"} 195.0
vllm:e2e_request_latency_seconds_bucket{le="15.0",model_name="study-model"} 195.0
vllm:e2e_request_latency_seconds_bucket{le="20.0",model_name="study-model"} 195.0
vllm:e2e_request_latency_seconds_bucket{le="30.0",model_name="study-model"} 195.0
vllm:e2e_request_latency_seconds_bucket{le="40.0",model_name="study-model"} 195.0
vllm:e2e_request_latency_seconds_bucket{le="50.0",model_name="study-model"} 195.0
vllm:e2e_request_latency_seconds_bucket{le="60.0",model_name="study-model"} 195.0
vllm:e2e_request_latency_seconds_bucket{le="+Inf",model_name="study-model"} 195.0
vllm:e2e_request_latency_seconds_count{model_name="study-model"} 195.0
vllm:e2e_request_latency_seconds_sum{model_name="study-model"} 648.9728918075562
# HELP vllm:request_prompt_tokens Number of prefill tokens processed.
# TYPE vllm:request_prompt_tokens histogram
vllm:request_prompt_tokens_bucket{le="1.0",model_name="study-model"} 0.0
vllm:request_prompt_tokens_bucket{le="2.0",model_name="study-model"} 0.0
vllm:request_prompt_tokens_bucket{le="5.0",model_name="study-model"} 0.0
vllm:request_prompt_tokens_bucket{le="10.0",model_name="study-model"} 0.0
vllm:request_prompt_tokens_bucket{le="20.0",model_name="study-model"} 0.0
vllm:request_prompt_tokens_bucket{le="50.0",model_name="study-model"} 0.0
vllm:request_prompt_tokens_bucket{le="100.0",model_name="study-model"} 0.0
vllm:request_prompt_tokens_bucket{le="200.0",model_name="study-model"} 147.0
vllm:request_prompt_tokens_bucket{le="500.0",model_name="study-model"} 147.0
vllm:request_prompt_tokens_bucket{le="1000.0",model_name="study-model"} 195.0
vllm:request_prompt_tokens_bucket{le="+Inf",model_name="study-model"} 195.0
vllm:request_prompt_tokens_count{model_name="study-model"} 195.0
vllm:request_prompt_tokens_sum{model_name="study-model"} 43392.0
# HELP vllm:request_generation_tokens Number of generation tokens processed.
# TYPE vllm:request_generation_tokens histogram
vllm:request_generation_tokens_bucket{le="1.0",model_name="study-model"} 0.0
vllm:request_generation_tokens_bucket{le="2.0",model_name="study-model"} 0.0
vllm:request_generation_tokens_bucket{le="5.0",model_name="study-model"} 0.0
vllm:request_generation_tokens_bucket{le="10.0",model_name="study-model"} 0.0
vllm:request_generation_tokens_bucket{le="20.0",model_name="study-model"} 0.0
vllm:request_generation_tokens_bucket{le="50.0",model_name="study-model"} 0.0
vllm:request_generation_tokens_bucket{le="100.0",model_name="study-model"} 195.0
vllm:request_generation_tokens_bucket{le="200.0",model_name="study-model"} 195.0
vllm:request_generation_tokens_bucket{le="500.0",model_name="study-model"} 195.0
vllm:request_generation_tokens_bucket{le="1000.0",model_name="study-model"} 195.0
vllm:request_generation_tokens_bucket{le="+Inf",model_name="study-model"} 195.0
vllm:request_generation_tokens_count{model_name="study-model"} 195.0
vllm:request_generation_tokens_sum{model_name="study-model"} 12480.0
# HELP vllm:request_params_best_of Histogram of the best_of request parameter.
# TYPE vllm:request_params_best_of histogram
vllm:request_params_best_of_bucket{le="1.0",model_name="study-model"} 195.0
vllm:request_params_best_of_bucket{le="2.0",model_name="study-model"} 195.0
vllm:request_params_best_of_bucket{le="5.0",model_name="study-model"} 195.0
vllm:request_params_best_of_bucket{le="10.0",model_name="study-model"} 195.0
vllm:request_params_best_of_bucket{le="20.0",model_name="study-model"} 195.0
vllm:request_params_best_of_bucket{le="+Inf",model_name="study-model"} 195.0
vllm:request_params_best_of_count{model_name="study-model"} 195.0
vllm:request_params_best_of_sum{model_name="study-model"} 195.0
# HELP vllm:request_params_n Histogram of the n request parameter.
# TYPE vllm:request_params_n histogram
vllm:request_params_n_bucket{le="1.0",model_name="study-model"} 195.0
vllm:request_params_n_bucket{le="2.0",model_name="study-model"} 195.0
vllm:request_params_n_bucket{le="5.0",model_name="study-model"} 195.0
vllm:request_params_n_bucket{le="10.0",model_name="study-model"} 195.0
vllm:request_params_n_bucket{le="20.0",model_name="study-model"} 195.0
vllm:request_params_n_bucket{le="+Inf",model_name="study-model"} 195.0
vllm:request_params_n_count{model_name="study-model"} 195.0
vllm:request_params_n_sum{model_name="study-model"} 195.0
# HELP vllm:request_success_total Count of successfully processed requests.
# TYPE vllm:request_success_total counter
vllm:request_success_total{finished_reason="length",model_name="study-model"} 195.0
# HELP vllm:spec_decode_draft_acceptance_rate Speulative token acceptance rate.
# TYPE vllm:spec_decode_draft_acceptance_rate gauge
# HELP vllm:spec_decode_efficiency Speculative decoding system efficiency.
# TYPE vllm:spec_decode_efficiency gauge
# HELP vllm:spec_decode_num_accepted_tokens_total Number of accepted tokens.
# TYPE vllm:spec_decode_num_accepted_tokens_total counter
# HELP vllm:spec_decode_num_draft_tokens_total Number of draft tokens.
# TYPE vllm:spec_decode_num_draft_tokens_total counter
# HELP vllm:spec_decode_num_emitted_tokens_total Number of emitted tokens.
# TYPE vllm:spec_decode_num_emitted_tokens_total counter
# HELP vllm:avg_prompt_throughput_toks_per_s Average prefill throughput in tokens/s.
# TYPE vllm:avg_prompt_throughput_toks_per_s gauge
vllm:avg_prompt_throughput_toks_per_s{model_name="study-model"} 0.0
# HELP vllm:avg_generation_throughput_toks_per_s Average generation throughput in tokens/s.
# TYPE vllm:avg_generation_throughput_toks_per_s gauge
vllm:avg_generation_throughput_toks_per_s{model_name="study-model"} 0.0
# HELP vllm:cache_config_info Information of the LLMEngine CacheConfig
# TYPE vllm:cache_config_info gauge
vllm:cache_config_info{block_size="16",cache_dtype="auto",cpu_offload_gb="0",enable_prefix_caching="False",gpu_memory_utilization="0.8",num_cpu_blocks="0",num_gpu_blocks="9815",num_gpu_blocks_override="None",sliding_window="None",swap_space_bytes="0.0"} 1.0
