# Same scrape as t1, OpenMetrics underscore names (some exporters rewrite ":").
vllm_num_requests_running{model_name="meta-llama/Llama-3.1-8B-Instruct"} 12.0
vllm_num_requests_waiting{model_name="meta-llama/Llama-3.1-8B-Instruct"} 0.0
vllm_kv_cache_usage_perc{model_name="meta-llama/Llama-3.1-8B-Instruct"} 0.38
vllm_prefix_cache_hits{model_name="meta-llama/Llama-3.1-8B-Instruct"} 610.0
vllm_prefix_cache_queries{model_name="meta-llama/Llama-3.1-8B-Instruct"} 1000.0
vllm_prompt_tokens_total{model_name="meta-llama/Llama-3.1-8B-Instruct"} 10210.0
vllm_generation_tokens_total{model_name="meta-llama/Llama-3.1-8B-Instruct"} 21632.0
vllm_time_to_first_token_seconds_sum{model_name="meta-llama/Llama-3.1-8B-Instruct"} 14.2
vllm_time_to_first_token_seconds_count{model_name="meta-llama/Llama-3.1-8B-Instruct"} 100.0
vllm_inter_token_latency_seconds_sum{model_name="meta-llama/Llama-3.1-8B-Instruct"} 3.1
vllm_inter_token_latency_seconds_count{model_name="meta-llama/Llama-3.1-8B-Instruct"} 100.0
vllm_e2e_request_latency_seconds_bucket{le="0.1",model_name="meta-llama/Llama-3.1-8B-Instruct"} 5.0
vllm_e2e_request_latency_seconds_bucket{le="0.25",model_name="meta-llama/Llama-3.1-8B-Instruct"} 20.0
vllm_e2e_request_latency_seconds_bucket{le="0.5",model_name="meta-llama/Llama-3.1-8B-Instruct"} 80.0
vllm_e2e_request_latency_seconds_bucket{le="1.0",model_name="meta-llama/Llama-3.1-8B-Instruct"} 95.0
vllm_e2e_request_latency_seconds_bucket{le="+Inf",model_name="meta-llama/Llama-3.1-8B-Instruct"} 100.0
vllm_e2e_request_latency_seconds_sum{model_name="meta-llama/Llama-3.1-8B-Instruct"} 42.0
vllm_e2e_request_latency_seconds_count{model_name="meta-llama/Llama-3.1-8B-Instruct"} 100.0
vllm_request_prompt_tokens_sum{model_name="meta-llama/Llama-3.1-8B-Instruct"} 800000.0
vllm_request_prompt_tokens_count{model_name="meta-llama/Llama-3.1-8B-Instruct"} 100.0
vllm_request_generation_tokens_sum{model_name="meta-llama/Llama-3.1-8B-Instruct"} 50000.0
vllm_request_generation_tokens_count{model_name="meta-llama/Llama-3.1-8B-Instruct"} 100.0
