# HELP sglang:process_cpu_seconds_total Total CPU time consumed by this process (user + system)
# TYPE sglang:process_cpu_seconds_total counter
sglang:process_cpu_seconds_total{component="tokenizer"} 5.000000000000002
sglang:process_cpu_seconds_total{component="detokenizer"} 0.09000000000000163
# HELP sglang:http_requests_total Total number of HTTP requests by endpoint and method
# TYPE sglang:http_requests_total counter
sglang:http_requests_total{endpoint="/model_info",method="GET"} 2.0
sglang:http_requests_total{endpoint="/v1/chat/completions",method="POST"} 2.0
sglang:http_requests_total{endpoint="/metrics",method="GET"} 1.0
# HELP sglang:http_responses_total Total number of HTTP responses by endpoint and status code
# TYPE sglang:http_responses_total counter
sglang:http_responses_total{endpoint="/model_info",method="GET",status_code="200"} 2.0
sglang:http_responses_total{endpoint="/v1/chat/completions",method="POST",status_code="200"} 2.0
# HELP sglang:prompt_tokens_total Number of prefill tokens processed.
# TYPE sglang:prompt_tokens_total counter
sglang:prompt_tokens_total{model_name="Qwen/Qwen3.5-35B-A3B-FP8"} 92.0
# HELP sglang:generation_tokens_total Number of generation tokens processed.
# TYPE sglang:generation_tokens_total counter
sglang:generation_tokens_total{model_name="Qwen/Qwen3.5-35B-A3B-FP8"} 9.0
# HELP sglang:num_requests_total Number of requests processed.
# TYPE sglang:num_requests_total counter
sglang:num_requests_total{model_name="Qwen/Qwen3.5-35B-A3B-FP8"} 2.0
# HELP sglang:cache_config_info Cache configuration information.
# TYPE sglang:cache_config_info gauge
sglang:cache_config_info{num_pages="788766",page_size="1"} 1.0
# HELP sglang:num_running_reqs The number of running requests.
# TYPE sglang:num_running_reqs gauge
sglang:num_running_reqs{engine_type="unified",model_name="Qwen/Qwen3.5-35B-A3B-FP8",moe_ep_rank="0",pp_rank="0",tp_rank="0"} 0.0
# HELP sglang:num_used_tokens The number of used tokens.
# TYPE sglang:num_used_tokens gauge
sglang:num_used_tokens{engine_type="unified",model_name="Qwen/Qwen3.5-35B-A3B-FP8",moe_ep_rank="0",pp_rank="0",tp_rank="0"} 0.0
# HELP sglang:token_usage The token usage.
# TYPE sglang:token_usage gauge
sglang:token_usage{engine_type="unified",model_name="Qwen/Qwen3.5-35B-A3B-FP8",moe_ep_rank="0",pp_rank="0",tp_rank="0"} 0.0
# HELP sglang:full_token_usage The token usage for full attention layers.
# TYPE sglang:full_token_usage gauge
sglang:full_token_usage{engine_type="unified",model_name="Qwen/Qwen3.5-35B-A3B-FP8",moe_ep_rank="0",pp_rank="0",tp_rank="0"} 0.00010142425003106117
# HELP sglang:pending_prealloc_token_usage The token usage for pending preallocated tokens (not preallocated yet).
# TYPE sglang:pending_prealloc_token_usage gauge
sglang:pending_prealloc_token_usage{engine_type="unified",model_name="Qwen/Qwen3.5-35B-A3B-FP8",moe_ep_rank="0",pp_rank="0",tp_rank="0"} 0.0
# HELP sglang:swa_token_usage The token usage for SWA layers.
# TYPE sglang:swa_token_usage gauge
sglang:swa_token_usage{engine_type="unified",model_name="Qwen/Qwen3.5-35B-A3B-FP8",moe_ep_rank="0",pp_rank="0",tp_rank="0"} 0.0
# HELP sglang:mamba_usage The token usage for Mamba layers.
# TYPE sglang:mamba_usage gauge
sglang:mamba_usage{engine_type="unified",model_name="Qwen/Qwen3.5-35B-A3B-FP8",moe_ep_rank="0",pp_rank="0",tp_rank="0"} 0.008888888888888889
# HELP sglang:decode_sum_seq_lens The sum of all sequence lengths in decode.
# TYPE sglang:decode_sum_seq_lens gauge
sglang:decode_sum_seq_lens{engine_type="unified",model_name="Qwen/Qwen3.5-35B-A3B-FP8",moe_ep_rank="0",pp_rank="0",tp_rank="0"} 0.0
# HELP sglang:gen_throughput The generation throughput (token/s).
# TYPE sglang:gen_throughput gauge
sglang:gen_throughput{engine_type="unified",model_name="Qwen/Qwen3.5-35B-A3B-FP8",moe_ep_rank="0",pp_rank="0",tp_rank="0"} 0.0
# HELP sglang:num_queue_reqs The number of requests in the waiting queue.
# TYPE sglang:num_queue_reqs gauge
sglang:num_queue_reqs{engine_type="unified",model_name="Qwen/Qwen3.5-35B-A3B-FP8",moe_ep_rank="0",pp_rank="0",tp_rank="0"} 0.0
# HELP sglang:num_grammar_queue_reqs The number of requests in the grammar waiting queue.
# TYPE sglang:num_grammar_queue_reqs gauge
sglang:num_grammar_queue_reqs{engine_type="unified",model_name="Qwen/Qwen3.5-35B-A3B-FP8",moe_ep_rank="0",pp_rank="0",tp_rank="0"} 0.0
# HELP sglang:num_running_reqs_offline_batch The number of running low-priority offline batch requests(label is 'batch').
# TYPE sglang:num_running_reqs_offline_batch gauge
sglang:num_running_reqs_offline_batch{engine_type="unified",model_name="Qwen/Qwen3.5-35B-A3B-FP8",moe_ep_rank="0",pp_rank="0",tp_rank="0"} 0.0
# HELP sglang:cache_hit_rate The prefix cache hit rate.
# TYPE sglang:cache_hit_rate gauge
sglang:cache_hit_rate{engine_type="unified",model_name="Qwen/Qwen3.5-35B-A3B-FP8",moe_ep_rank="0",pp_rank="0",tp_rank="0"} 0.0
# HELP sglang:max_total_num_tokens Maximum total number of tokens in the KV cache pool.
# TYPE sglang:max_total_num_tokens gauge
sglang:max_total_num_tokens{engine_type="unified",model_name="Qwen/Qwen3.5-35B-A3B-FP8",moe_ep_rank="0",pp_rank="0",tp_rank="0"} 788766.0
# HELP sglang:spec_accept_length The average acceptance length of speculative decoding.
# TYPE sglang:spec_accept_length gauge
sglang:spec_accept_length{engine_type="unified",model_name="Qwen/Qwen3.5-35B-A3B-FP8",moe_ep_rank="0",pp_rank="0",tp_rank="0"} 0.0
# HELP sglang:spec_accept_rate The average acceptance rate of speculative decoding (`accepted tokens / total draft tokens` in batch).
# TYPE sglang:spec_accept_rate gauge
sglang:spec_accept_rate{engine_type="unified",model_name="Qwen/Qwen3.5-35B-A3B-FP8",moe_ep_rank="0",pp_rank="0",tp_rank="0"} 0.0
# HELP sglang:num_prefill_prealloc_queue_reqs The number of requests in the prefill prealloc queue.
# TYPE sglang:num_prefill_prealloc_queue_reqs gauge
sglang:num_prefill_prealloc_queue_reqs{engine_type="unified",model_name="Qwen/Qwen3.5-35B-A3B-FP8",moe_ep_rank="0",pp_rank="0",tp_rank="0"} 0.0
# HELP sglang:num_prefill_inflight_queue_reqs The number of requests in the prefill inflight queue.
# TYPE sglang:num_prefill_inflight_queue_reqs gauge
sglang:num_prefill_inflight_queue_reqs{engine_type="unified",model_name="Qwen/Qwen3.5-35B-A3B-FP8",moe_ep_rank="0",pp_rank="0",tp_rank="0"} 0.0
# HELP sglang:num_decode_prealloc_queue_reqs The number of requests in the decode prealloc queue.
# TYPE sglang:num_decode_prealloc_queue_reqs gauge
sglang:num_decode_prealloc_queue_reqs{engine_type="unified",model_name="Qwen/Qwen3.5-35B-A3B-FP8",moe_ep_rank="0",pp_rank="0",tp_rank="0"} 0.0
# HELP sglang:num_decode_transfer_queue_reqs The number of requests in the decode transfer queue.
# TYPE sglang:num_decode_transfer_queue_reqs gauge
sglang:num_decode_transfer_queue_reqs{engine_type="unified",model_name="Qwen/Qwen3.5-35B-A3B-FP8",moe_ep_rank="0",pp_rank="0",tp_rank="0"} 0.0
# HELP sglang:utilization The utilization.
# TYPE sglang:utilization gauge
sglang:utilization{engine_type="unified",model_name="Qwen/Qwen3.5-35B-A3B-FP8",moe_ep_rank="0",pp_rank="0",tp_rank="0"} 0.0
# HELP sglang:engine_startup_time The time taken for the engine to start up.
# TYPE sglang:engine_startup_time gauge
sglang:engine_startup_time{engine_type="unified",model_name="Qwen/Qwen3.5-35B-A3B-FP8",moe_ep_rank="0",pp_rank="0",tp_rank="0"} 0.0
# HELP sglang:engine_load_weights_time The time taken for the engine to load weights.
# TYPE sglang:engine_load_weights_time gauge
sglang:engine_load_weights_time{engine_type="unified",model_name="Qwen/Qwen3.5-35B-A3B-FP8",moe_ep_rank="0",pp_rank="0",tp_rank="0"} 0.0
# HELP sglang:new_token_ratio The new token ratio.
# TYPE sglang:new_token_ratio gauge
sglang:new_token_ratio{engine_type="unified",model_name="Qwen/Qwen3.5-35B-A3B-FP8",moe_ep_rank="0",pp_rank="0",tp_rank="0"} 0.7
# HELP sglang:is_cuda_graph Whether the batch is using CUDA graph.
# TYPE sglang:is_cuda_graph gauge
sglang:is_cuda_graph{engine_type="unified",model_name="Qwen/Qwen3.5-35B-A3B-FP8",moe_ep_rank="0",pp_rank="0",tp_rank="0"} 0.0
# HELP sglang:num_unique_running_routing_keys Number of unique routing keys in running batch.
# TYPE sglang:num_unique_running_routing_keys gauge
sglang:num_unique_running_routing_keys{engine_type="unified",model_name="Qwen/Qwen3.5-35B-A3B-FP8",moe_ep_rank="0",pp_rank="0",tp_rank="0"} 0.0
# HELP sglang:routing_key_running_req_count Distribution of routing keys by running request count (gt < count <= le).
# TYPE sglang:routing_key_running_req_count gauge
sglang:routing_key_running_req_count{engine_type="unified",gt="0",le="1",model_name="Qwen/Qwen3.5-35B-A3B-FP8",moe_ep_rank="0",pp_rank="0",tp_rank="0"} 0.0
sglang:routing_key_running_req_count{engine_type="unified",gt="1",le="2",model_name="Qwen/Qwen3.5-35B-A3B-FP8",moe_ep_rank="0",pp_rank="0",tp_rank="0"} 0.0
sglang:routing_key_running_req_count{engine_type="unified",gt="2",le="3",model_name="Qwen/Qwen3.5-35B-A3B-FP8",moe_ep_rank="0",pp_rank="0",tp_rank="0"} 0.0
sglang:routing_key_running_req_count{engine_type="unified",gt="3",le="5",model_name="Qwen/Qwen3.5-35B-A3B-FP8",moe_ep_rank="0",pp_rank="0",tp_rank="0"} 0.0
sglang:routing_key_running_req_count{engine_type="unified",gt="5",le="7",model_name="Qwen/Qwen3.5-35B-A3B-FP8",moe_ep_rank="0",pp_rank="0",tp_rank="0"} 0.0
sglang:routing_key_running_req_count{engine_type="unified",gt="7",le="10",model_name="Qwen/Qwen3.5-35B-A3B-FP8",moe_ep_rank="0",pp_rank="0",tp_rank="0"} 0.0
sglang:routing_key_running_req_count{engine_type="unified",gt="10",le="20",model_name="Qwen/Qwen3.5-35B-A3B-FP8",moe_ep_rank="0",pp_rank="0",tp_rank="0"} 0.0
sglang:routing_key_running_req_count{engine_type="unified",gt="20",le="50",model_name="Qwen/Qwen3.5-35B-A3B-FP8",moe_ep_rank="0",pp_rank="0",tp_rank="0"} 0.0
sglang:routing_key_running_req_count{engine_type="unified",gt="50",le="100",model_name="Qwen/Qwen3.5-35B-A3B-FP8",moe_ep_rank="0",pp_rank="0",tp_rank="0"} 0.0
sglang:routing_key_running_req_count{engine_type="unified",gt="100",le="200",model_name="Qwen/Qwen3.5-35B-A3B-FP8",moe_ep_rank="0",pp_rank="0",tp_rank="0"} 0.0
sglang:routing_key_running_req_count{engine_type="unified",gt="200",le="+Inf",model_name="Qwen/Qwen3.5-35B-A3B-FP8",moe_ep_rank="0",pp_rank="0",tp_rank="0"} 0.0
# HELP sglang:routing_key_all_req_count Distribution of routing keys by running+waiting request count (gt < count <= le).
# TYPE sglang:routing_key_all_req_count gauge
sglang:routing_key_all_req_count{engine_type="unified",gt="0",le="1",model_name="Qwen/Qwen3.5-35B-A3B-FP8",moe_ep_rank="0",pp_rank="0",tp_rank="0"} 0.0
sglang:routing_key_all_req_count{engine_type="unified",gt="1",le="2",model_name="Qwen/Qwen3.5-35B-A3B-FP8",moe_ep_rank="0",pp_rank="0",tp_rank="0"} 0.0
sglang:routing_key_all_req_count{engine_type="unified",gt="2",le="3",model_name="Qwen/Qwen3.5-35B-A3B-FP8",moe_ep_rank="0",pp_rank="0",tp_rank="0"} 0.0
sglang:routing_key_all_req_count{engine_type="unified",gt="3",le="5",model_name="Qwen/Qwen3.5-35B-A3B-FP8",moe_ep_rank="0",pp_rank="0",tp_rank="0"} 0.0
sglang:routing_key_all_req_count{engine_type="unified",gt="5",le="7",model_name="Qwen/Qwen3.5-35B-A3B-FP8",moe_ep_rank="0",pp_rank="0",tp_rank="0"} 0.0
sglang:routing_key_all_req_count{engine_type="unified",gt="7",le="10",model_name="Qwen/Qwen3.5-35B-A3B-FP8",moe_ep_rank="0",pp_rank="0",tp_rank="0"} 0.0
sglang:routing_key_all_req_count{engine_type="unified",gt="10",le="20",model_name="Qwen/Qwen3.5-35B-A3B-FP8",moe_ep_rank="0",pp_rank="0",tp_rank="0"} 0.0
sglang:routing_key_all_req_count{engine_type="unified",gt="20",le="50",model_name="Qwen/Qwen3.5-35B-A3B-FP8",moe_ep_rank="0",pp_rank="0",tp_rank="0"} 0.0
sglang:routing_key_all_req_count{engine_type="unified",gt="50",le="100",model_name="Qwen/Qwen3.5-35B-A3B-FP8",moe_ep_rank="0",pp_rank="0",tp_rank="0"} 0.0
sglang:routing_key_all_req_count{engine_type="unified",gt="100",le="200",model_name="Qwen/Qwen3.5-35B-A3B-FP8",moe_ep_rank="0",pp_rank="0",tp_rank="0"} 0.0
sglang:routing_key_all_req_count{engine_type="unified",gt="200",le="+Inf",model_name="Qwen/Qwen3.5-35B-A3B-FP8",moe_ep_rank="0",pp_rank="0",tp_rank="0"} 0.0
# HELP sglang:routing_keys_active Number of unique routing keys with active requests
# TYPE sglang:routing_keys_active gauge
sglang:routing_keys_active 0.0
# HELP sglang:http_requests_active Number of currently active HTTP requests
# TYPE sglang:http_requests_active gauge
sglang:http_requests_active{endpoint="/model_info",method="GET"} 0.0
sglang:http_requests_active{endpoint="/v1/chat/completions",method="POST"} 0.0
sglang:http_requests_active{endpoint="/metrics",method="GET"} 1.0
# HELP sglang:num_retracted_reqs The number of retracted requests.
# TYPE sglang:num_retracted_reqs gauge
sglang:num_retracted_reqs{engine_type="unified",model_name="Qwen/Qwen3.5-35B-A3B-FP8",moe_ep_rank="0",pid="21",pp_rank="0",tp_rank="0"} 0.0
# HELP sglang:num_paused_reqs The number of paused requests by async weight sync.
# TYPE sglang:num_paused_reqs gauge
sglang:num_paused_reqs{engine_type="unified",model_name="Qwen/Qwen3.5-35B-A3B-FP8",moe_ep_rank="0",pid="21",pp_rank="0",tp_rank="0"} 0.0
# HELP sglang:per_stage_req_latency_seconds The latency of each stage of requests.
# TYPE sglang:per_stage_req_latency_seconds histogram
sglang:per_stage_req_latency_seconds_sum{engine_type="unified",model_name="Qwen/Qwen3.5-35B-A3B-FP8",moe_ep_rank="0",pp_rank="0",stage="request_process",tp_rank="0"} 0.0016043189999948027
sglang:per_stage_req_latency_seconds_bucket{engine_type="unified",le="0.001",model_name="Qwen/Qwen3.5-35B-A3B-FP8",moe_ep_rank="0",pp_rank="0",stage="request_process",tp_rank="0"} 1.0
sglang:per_stage_req_latency_seconds_bucket{engine_type="unified",le="0.0016200000000000001",model_name="Qwen/Qwen3.5-35B-A3B-FP8",moe_ep_rank="0",pp_rank="0",stage="request_process",tp_rank="0"} 2.0
sglang:per_stage_req_latency_seconds_bucket{engine_type="unified",le="0.0026244000000000007",model_name="Qwen/Qwen3.5-35B-A3B-FP8",moe_ep_rank="0",pp_rank="0",stage="request_process",tp_rank="0"} 2.0
sglang:per_stage_req_latency_seconds_bucket{engine_type="unified",le="0.004251528",model_name="Qwen/Qwen3.5-35B-A3B-FP8",moe_ep_rank="0",pp_rank="0",stage="request_process",tp_rank="0"} 2.0
sglang:per_stage_req_latency_seconds_bucket{engine_type="unified",le="0.006887475360000002",model_name="Qwen/Qwen3.5-35B-A3B-FP8",moe_ep_rank="0",pp_rank="0",stage="request_process",tp_rank="0"} 2.0
sglang:per_stage_req_latency_seconds_bucket{engine_type="unified",le="0.011157710083200004",model_name="Qwen/Qwen3.5-35B-A3B-FP8",moe_ep_rank="0",pp_rank="0",stage="request_process",tp_rank="0"} 2.0
sglang:per_stage_req_latency_seconds_bucket{engine_type="unified",le="0.018075490334784006",model_name="Qwen/Qwen3.5-35B-A3B-FP8",moe_ep_rank="0",pp_rank="0",stage="request_process",tp_rank="0"} 2.0
sglang:per_stage_req_latency_seconds_bucket{engine_type="unified",le="0.029282294342350097",model_name="Qwen/Qwen3.5-35B-A3B-FP8",moe_ep_rank="0",pp_rank="0",stage="request_process",tp_rank="0"} 2.0
sglang:per_stage_req_latency_seconds_bucket{engine_type="unified",le="0.047437316834607154",model_name="Qwen/Qwen3.5-35B-A3B-FP8",moe_ep_rank="0",pp_rank="0",stage="request_process",tp_rank="0"} 2.0
sglang:per_stage_req_latency_seconds_bucket{engine_type="unified",le="0.07684845327206359",model_name="Qwen/Qwen3.5-35B-A3B-FP8",moe_ep_rank="0",pp_rank="0",stage="request_process",tp_rank="0"} 2.0
sglang:per_stage_req_latency_seconds_bucket{engine_type="unified",le="0.12449449430074302",model_name="Qwen/Qwen3.5-35B-A3B-FP8",moe_ep_rank="0",pp_rank="0",stage="request_process",tp_rank="0"} 2.0
sglang:per_stage_req_latency_seconds_bucket{engine_type="unified",le="0.20168108076720373",model_name="Qwen/Qwen3.5-35B-A3B-FP8",moe_ep_rank="0",pp_rank="0",stage="request_process",tp_rank="0"} 2.0
sglang:per_stage_req_latency_seconds_bucket{engine_type="unified",le="0.3267233508428701",model_name="Qwen/Qwen3.5-35B-A3B-FP8",moe_ep_rank="0",pp_rank="0",stage="request_process",tp_rank="0"} 2.0
sglang:per_stage_req_latency_seconds_bucket{engine_type="unified",le="0.5292918283654495",model_name="Qwen/Qwen3.5-35B-A3B-FP8",moe_ep_rank="0",pp_rank="0",stage="request_process",tp_rank="0"} 2.0
sglang:per_stage_req_latency_seconds_bucket{engine_type="unified",le="0.8574527619520282",model_name="Qwen/Qwen3.5-35B-A3B-FP8",moe_ep_rank="0",pp_rank="0",stage="request_process",tp_rank="0"} 2.0
sglang:per_stage_req_latency_seconds_bucket{engine_type="unified",le="1.389073474362286",model_name="Qwen/Qwen3.5-35B-A3B-FP8",moe_ep_rank="0",pp_rank="0",stage="request_process",tp_rank="0"} 2.0
sglang:per_stage_req_latency_seconds_bucket{engine_type="unified",le="2.2502990284669036",model_name="Qwen/Qwen3.5-35B-A3B-FP8",moe_ep_rank="0",pp_rank="0",stage="request_process",tp_rank="0"} 2.0
sglang:per_stage_req_latency_seconds_bucket{engine_type="unified",le="3.6454844261163837",model_name="Qwen/Qwen3.5-35B-A3B-FP8",moe_ep_rank="0",pp_rank="0",stage="request_process",tp_rank="0"} 2.0
sglang:per_stage_req_latency_seconds_bucket{engine_type="unified",le="5.905684770308542",model_name="Qwen/Qwen3.5-35B-A3B-FP8",moe_ep_rank="0",pp_rank="0",stage="request_process",tp_rank="0"} 2.0
sglang:per_stage_req_latency_seconds_bucket{engine_type="unified",le="9.567209327899839",model_name="Qwen/Qwen3.5-35B-A3B-FP8",moe_ep_rank="0",pp_rank="0",stage="request_process",tp_rank="0"} 2.0
sglang:per_stage_req_latency_seconds_bucket{engine_type="unified",le="15.49887911119774",model_name="Qwen/Qwen3.5-35B-A3B-FP8",moe_ep_rank="0",pp_rank="0",stage="request_process",tp_rank="0"} 2.0
sglang:per_stage_req_latency_seconds_bucket{engine_type="unified",le="25.10818416014034",model_name="Qwen/Qwen3.5-35B-A3B-FP8",moe_ep_rank="0",pp_rank="0",stage="request_process",tp_rank="0"} 2.0
sglang:per_stage_req_latency_seconds_bucket{engine_type="unified",le="40.67525833942735",model_name="Qwen/Qwen3.5-35B-A3B-FP8",moe_ep_rank="0",pp_rank="0",stage="request_process",tp_rank="0"} 2.0
sglang:per_stage_req_latency_seconds_bucket{engine_type="unified",le="65.89391850987231",model_name="Qwen/Qwen3.5-35B-A3B-FP8",moe_ep_rank="0",pp_rank="0",stage="request_process",tp_rank="0"} 2.0
sglang:per_stage_req_latency_seconds_bucket{engine_type="unified",le="106.74814798599316",model_name="Qwen/Qwen3.5-35B-A3B-FP8",moe_ep_rank="0",pp_rank="0",stage="request_process",tp_rank="0"} 2.0
sglang:per_stage_req_latency_seconds_bucket{engine_type="unified",le="172.93199973730896",model_name="Qwen/Qwen3.5-35B-A3B-FP8",moe_ep_rank="0",pp_rank="0",stage="request_process",tp_rank="0"} 2.0
sglang:per_stage_req_latency_seconds_bucket{engine_type="unified",le="280.1498395744405",model_name="Qwen/Qwen3.5-35B-A3B-FP8",moe_ep_rank="0",pp_rank="0",stage="request_process",tp_rank="0"} 2.0
sglang:per_stage_req_latency_seconds_bucket{engine_type="unified",le="453.84274011059364",model_name="Qwen/Qwen3.5-35B-A3B-FP8",moe_ep_rank="0",pp_rank="0",stage="request_process",tp_rank="0"} 2.0
sglang:per_stage_req_latency_seconds_bucket{engine_type="unified",le="735.2252389791618",model_name="Qwen/Qwen3.5-35B-A3B-FP8",moe_ep_rank="0",pp_rank="0",stage="request_process",tp_rank="0"} 2.0
sglang:per_stage_req_latency_seconds_bucket{engine_type="unified",le="1191.064887146242",model_name="Qwen/Qwen3.5-35B-A3B-FP8",moe_ep_rank="0",pp_rank="0",stage="request_process",tp_rank="0"} 2.0
sglang:per_stage_req_latency_seconds_bucket{engine_type="unified",le="+Inf",model_name="Qwen/Qwen3.5-35B-A3B-FP8",moe_ep_rank="0",pp_rank="0",stage="request_process",tp_rank="0"} 2.0
sglang:per_stage_req_latency_seconds_count{engine_type="unified",model_name="Qwen/Qwen3.5-35B-A3B-FP8",moe_ep_rank="0",pp_rank="0",stage="request_process",tp_rank="0"} 2.0
sglang:per_stage_req_latency_seconds_sum{engine_type="unified",model_name="Qwen/Qwen3.5-35B-A3B-FP8",moe_ep_rank="0",pp_rank="0",stage="prefill_forward",tp_rank="0"} 18.316858581999995
sglang:per_stage_req_latency_seconds_bucket{engine_type="unified",le="0.001",model_name="Qwen/Qwen3.5-35B-A3B-FP8",moe_ep_rank="0",pp_rank="0",stage="prefill_forward",tp_rank="0"} 0.0
sglang:per_stage_req_latency_seconds_bucket{engine_type="unified",le="0.0016200000000000001",model_name="Qwen/Qwen3.5-35B-A3B-FP8",moe_ep_rank="0",pp_rank="0",stage="prefill_forward",tp_rank="0"} 0.0
sglang:per_stage_req_latency_seconds_bucket{engine_type="unified",le="0.0026244000000000007",model_name="Qwen/Qwen3.5-35B-A3B-FP8",moe_ep_rank="0",pp_rank="0",stage="prefill_forward",tp_rank="0"} 0.0
sglang:per_stage_req_latency_seconds_bucket{engine_type="unified",le="0.004251528",model_name="Qwen/Qwen3.5-35B-A3B-FP8",moe_ep_rank="0",pp_rank="0",stage="prefill_forward",tp_rank="0"} 0.0
sglang:per_stage_req_latency_seconds_bucket{engine_type="unified",le="0.006887475360000002",model_name="Qwen/Qwen3.5-35B-A3B-FP8",moe_ep_rank="0",pp_rank="0",stage="prefill_forward",tp_rank="0"} 0.0
sglang:per_stage_req_latency_seconds_bucket{engine_type="unified",le="0.011157710083200004",model_name="Qwen/Qwen3.5-35B-A3B-FP8",moe_ep_rank="0",pp_rank="0",stage="prefill_forward",tp_rank="0"} 0.0
sglang:per_stage_req_latency_seconds_bucket{engine_type="unified",le="0.018075490334784006",model_name="Qwen/Qwen3.5-35B-A3B-FP8",moe_ep_rank="0",pp_rank="0",stage="prefill_forward",tp_rank="0"} 0.0
sglang:per_stage_req_latency_seconds_bucket{engine_type="unified",le="0.029282294342350097",model_name="Qwen/Qwen3.5-35B-A3B-FP8",moe_ep_rank="0",pp_rank="0",stage="prefill_forward",tp_rank="0"} 0.0
sglang:per_stage_req_latency_seconds_bucket{engine_type="unified",le="0.047437316834607154",model_name="Qwen/Qwen3.5-35B-A3B-FP8",moe_ep_rank="0",pp_rank="0",stage="prefill_forward",tp_rank="0"} 0.0
sglang:per_stage_req_latency_seconds_bucket{engine_type="unified",le="0.07684845327206359",model_name="Qwen/Qwen3.5-35B-A3B-FP8",moe_ep_rank="0",pp_rank="0",stage="prefill_forward",tp_rank="0"} 0.0
sglang:per_stage_req_latency_seconds_bucket{engine_type="unified",le="0.12449449430074302",model_name="Qwen/Qwen3.5-35B-A3B-FP8",moe_ep_rank="0",pp_rank="0",stage="prefill_forward",tp_rank="0"} 0.0
sglang:per_stage_req_latency_seconds_bucket{engine_type="unified",le="0.20168108076720373",model_name="Qwen/Qwen3.5-35B-A3B-FP8",moe_ep_rank="0",pp_rank="0",stage="prefill_forward",tp_rank="0"} 0.0
sglang:per_stage_req_latency_seconds_bucket{engine_type="unified",le="0.3267233508428701",model_name="Qwen/Qwen3.5-35B-A3B-FP8",moe_ep_rank="0",pp_rank="0",stage="prefill_forward",tp_rank="0"} 0.0
sglang:per_stage_req_latency_seconds_bucket{engine_type="unified",le="0.5292918283654495",model_name="Qwen/Qwen3.5-35B-A3B-FP8",moe_ep_rank="0",pp_rank="0",stage="prefill_forward",tp_rank="0"} 0.0
sglang:per_stage_req_latency_seconds_bucket{engine_type="unified",le="0.8574527619520282",model_name="Qwen/Qwen3.5-35B-A3B-FP8",moe_ep_rank="0",pp_rank="0",stage="prefill_forward",tp_rank="0"} 0.0
sglang:per_stage_req_latency_seconds_bucket{engine_type="unified",le="1.389073474362286",model_name="Qwen/Qwen3.5-35B-A3B-FP8",moe_ep_rank="0",pp_rank="0",stage="prefill_forward",tp_rank="0"} 0.0
sglang:per_stage_req_latency_seconds_bucket{engine_type="unified",le="2.2502990284669036",model_name="Qwen/Qwen3.5-35B-A3B-FP8",moe_ep_rank="0",pp_rank="0",stage="prefill_forward",tp_rank="0"} 1.0
sglang:per_stage_req_latency_seconds_bucket{engine_type="unified",le="3.6454844261163837",model_name="Qwen/Qwen3.5-35B-A3B-FP8",moe_ep_rank="0",pp_rank="0",stage="prefill_forward",tp_rank="0"} 1.0
sglang:per_stage_req_latency_seconds_bucket{engine_type="unified",le="5.905684770308542",model_name="Qwen/Qwen3.5-35B-A3B-FP8",moe_ep_rank="0",pp_rank="0",stage="prefill_forward",tp_rank="0"} 1.0
sglang:per_stage_req_latency_seconds_bucket{engine_type="unified",le="9.567209327899839",model_name="Qwen/Qwen3.5-35B-A3B-FP8",moe_ep_rank="0",pp_rank="0",stage="prefill_forward",tp_rank="0"} 1.0
sglang:per_stage_req_latency_seconds_bucket{engine_type="unified",le="15.49887911119774",model_name="Qwen/Qwen3.5-35B-A3B-FP8",moe_ep_rank="0",pp_rank="0",stage="prefill_forward",tp_rank="0"} 1.0
sglang:per_stage_req_latency_seconds_bucket{engine_type="unified",le="25.10818416014034",model_name="Qwen/Qwen3.5-35B-A3B-FP8",moe_ep_rank="0",pp_rank="0",stage="prefill_forward",tp_rank="0"} 2.0
sglang:per_stage_req_latency_seconds_bucket{engine_type="unified",le="40.67525833942735",model_name="Qwen/Qwen3.5-35B-A3B-FP8",moe_ep_rank="0",pp_rank="0",stage="prefill_forward",tp_rank="0"} 2.0
sglang:per_stage_req_latency_seconds_bucket{engine_type="unified",le="65.89391850987231",model_name="Qwen/Qwen3.5-35B-A3B-FP8",moe_ep_rank="0",pp_rank="0",stage="prefill_forward",tp_rank="0"} 2.0
sglang:per_stage_req_latency_seconds_bucket{engine_type="unified",le="106.74814798599316",model_name="Qwen/Qwen3.5-35B-A3B-FP8",moe_ep_rank="0",pp_rank="0",stage="prefill_forward",tp_rank="0"} 2.0
sglang:per_stage_req_latency_seconds_bucket{engine_type="unified",le="172.93199973730896",model_name="Qwen/Qwen3.5-35B-A3B-FP8",moe_ep_rank="0",pp_rank="0",stage="prefill_forward",tp_rank="0"} 2.0
sglang:per_stage_req_latency_seconds_bucket{engine_type="unified",le="280.1498395744405",model_name="Qwen/Qwen3.5-35B-A3B-FP8",moe_ep_rank="0",pp_rank="0",stage="prefill_forward",tp_rank="0"} 2.0
sglang:per_stage_req_latency_seconds_bucket{engine_type="unified",le="453.84274011059364",model_name="Qwen/Qwen3.5-35B-A3B-FP8",moe_ep_rank="0",pp_rank="0",stage="prefill_forward",tp_rank="0"} 2.0
sglang:per_stage_req_latency_seconds_bucket{engine_type="unified",le="735.2252389791618",model_name="Qwen/Qwen3.5-35B-A3B-FP8",moe_ep_rank="0",pp_rank="0",stage="prefill_forward",tp_rank="0"} 2.0
sglang:per_stage_req_latency_seconds_bucket{engine_type="unified",le="1191.064887146242",model_name="Qwen/Qwen3.5-35B-A3B-FP8",moe_ep_rank="0",pp_rank="0",stage="prefill_forward",tp_rank="0"} 2.0
sglang:per_stage_req_latency_seconds_bucket{engine_type="unified",le="+Inf",model_name="Qwen/Qwen3.5-35B-A3B-FP8",moe_ep_rank="0",pp_rank="0",stage="prefill_forward",tp_rank="0"} 2.0
sglang:per_stage_req_latency_seconds_count{engine_type="unified",model_name="Qwen/Qwen3.5-35B-A3B-FP8",moe_ep_rank="0",pp_rank="0",stage="prefill_forward",tp_rank="0"} 2.0
# HELP sglang:queue_time_seconds Histogram of queueing time in seconds.
# TYPE sglang:queue_time_seconds histogram
sglang:queue_time_seconds_sum{engine_type="unified",model_name="Qwen/Qwen3.5-35B-A3B-FP8",moe_ep_rank="0",pp_rank="0",tp_rank="0"} 0.0018174989999977242
sglang:queue_time_seconds_bucket{engine_type="unified",le="0.0",model_name="Qwen/Qwen3.5-35B-A3B-FP8",moe_ep_rank="0",pp_rank="0",tp_rank="0"} 0.0
sglang:queue_time_seconds_bucket{engine_type="unified",le="0.1",model_name="Qwen/Qwen3.5-35B-A3B-FP8",moe_ep_rank="0",pp_rank="0",tp_rank="0"} 2.0
sglang:queue_time_seconds_bucket{engine_type="unified",le="0.2",model_name="Qwen/Qwen3.5-35B-A3B-FP8",moe_ep_rank="0",pp_rank="0",tp_rank="0"} 2.0
sglang:queue_time_seconds_bucket{engine_type="unified",le="0.5",model_name="Qwen/Qwen3.5-35B-A3B-FP8",moe_ep_rank="0",pp_rank="0",tp_rank="0"} 2.0
sglang:queue_time_seconds_bucket{engine_type="unified",le="1.0",model_name="Qwen/Qwen3.5-35B-A3B-FP8",moe_ep_rank="0",pp_rank="0",tp_rank="0"} 2.0
sglang:queue_time_seconds_bucket{engine_type="unified",le="2.0",model_name="Qwen/Qwen3.5-35B-A3B-FP8",moe_ep_rank="0",pp_rank="0",tp_rank="0"} 2.0
sglang:queue_time_seconds_bucket{engine_type="unified",le="3.0",model_name="Qwen/Qwen3.5-35B-A3B-FP8",moe_ep_rank="0",pp_rank="0",tp_rank="0"} 2.0
sglang:queue_time_seconds_bucket{engine_type="unified",le="4.0",model_name="Qwen/Qwen3.5-35B-A3B-FP8",moe_ep_rank="0",pp_rank="0",tp_rank="0"} 2.0
sglang:queue_time_seconds_bucket{engine_type="unified",le="5.0",model_name="Qwen/Qwen3.5-35B-A3B-FP8",moe_ep_rank="0",pp_rank="0",tp_rank="0"} 2.0
sglang:queue_time_seconds_bucket{engine_type="unified",le="10.0",model_name="Qwen/Qwen3.5-35B-A3B-FP8",moe_ep_rank="0",pp_rank="0",tp_rank="0"} 2.0
sglang:queue_time_seconds_bucket{engine_type="unified",le="15.0",model_name="Qwen/Qwen3.5-35B-A3B-FP8",moe_ep_rank="0",pp_rank="0",tp_rank="0"} 2.0
sglang:queue_time_seconds_bucket{engine_type="unified",le="20.0",model_name="Qwen/Qwen3.5-35B-A3B-FP8",moe_ep_rank="0",pp_rank="0",tp_rank="0"} 2.0
sglang:queue_time_seconds_bucket{engine_type="unified",le="30.0",model_name="Qwen/Qwen3.5-35B-A3B-FP8",moe_ep_rank="0",pp_rank="0",tp_rank="0"} 2.0
sglang:queue_time_seconds_bucket{engine_type="unified",le="40.0",model_name="Qwen/Qwen3.5-35B-A3B-FP8",moe_ep_rank="0",pp_rank="0",tp_rank="0"} 2.0
sglang:queue_time_seconds_bucket{engine_type="unified",le="50.0",model_name="Qwen/Qwen3.5-35B-A3B-FP8",moe_ep_rank="0",pp_rank="0",tp_rank="0"} 2.0
sglang:queue_time_seconds_bucket{engine_type="unified",le="60.0",model_name="Qwen/Qwen3.5-35B-A3B-FP8",moe_ep_rank="0",pp_rank="0",tp_rank="0"} 2.0
sglang:queue_time_seconds_bucket{engine_type="unified",le="70.0",model_name="Qwen/Qwen3.5-35B-A3B-FP8",moe_ep_rank="0",pp_rank="0",tp_rank="0"} 2.0
sglang:queue_time_seconds_bucket{engine_type="unified",le="80.0",model_name="Qwen/Qwen3.5-35B-A3B-FP8",moe_ep_rank="0",pp_rank="0",tp_rank="0"} 2.0
sglang:queue_time_seconds_bucket{engine_type="unified",le="90.0",model_name="Qwen/Qwen3.5-35B-A3B-FP8",moe_ep_rank="0",pp_rank="0",tp_rank="0"} 2.0
sglang:queue_time_seconds_bucket{engine_type="unified",le="100.0",model_name="Qwen/Qwen3.5-35B-A3B-FP8",moe_ep_rank="0",pp_rank="0",tp_rank="0"} 2.0
sglang:queue_time_seconds_bucket{engine_type="unified",le="200.0",model_name="Qwen/Qwen3.5-35B-A3B-FP8",moe_ep_rank="0",pp_rank="0",tp_rank="0"} 2.0
sglang:queue_time_seconds_bucket{engine_type="unified",le="300.0",model_name="Qwen/Qwen3.5-35B-A3B-FP8",moe_ep_rank="0",pp_rank="0",tp_rank="0"} 2.0
sglang:queue_time_seconds_bucket{engine_type="unified",le="400.0",model_name="Qwen/Qwen3.5-35B-A3B-FP8",moe_ep_rank="0",pp_rank="0",tp_rank="0"} 2.0
sglang:queue_time_seconds_bucket{engine_type="unified",le="500.0",model_name="Qwen/Qwen3.5-35B-A3B-FP8",moe_ep_rank="0",pp_rank="0",tp_rank="0"} 2.0
sglang:queue_time_seconds_bucket{engine_type="unified",le="600.0",model_name="Qwen/Qwen3.5-35B-A3B-FP8",moe_ep_rank="0",pp_rank="0",tp_rank="0"} 2.0
sglang:queue_time_seconds_bucket{engine_type="unified",le="700.0",model_name="Qwen/Qwen3.5-35B-A3B-FP8",moe_ep_rank="0",pp_rank="0",tp_rank="0"} 2.0
sglang:queue_time_seconds_bucket{engine_type="unified",le="800.0",model_name="Qwen/Qwen3.5-35B-A3B-FP8",moe_ep_rank="0",pp_rank="0",tp_rank="0"} 2.0
sglang:queue_time_seconds_bucket{engine_type="unified",le="900.0",model_name="Qwen/Qwen3.5-35B-A3B-FP8",moe_ep_rank="0",pp_rank="0",tp_rank="0"} 2.0
sglang:queue_time_seconds_bucket{engine_type="unified",le="1000.0",model_name="Qwen/Qwen3.5-35B-A3B-FP8",moe_ep_rank="0",pp_rank="0",tp_rank="0"} 2.0
sglang:queue_time_seconds_bucket{engine_type="unified",le="1200.0",model_name="Qwen/Qwen3.5-35B-A3B-FP8",moe_ep_rank="0",pp_rank="0",tp_rank="0"} 2.0
sglang:queue_time_seconds_bucket{engine_type="unified",le="1400.0",model_name="Qwen/Qwen3.5-35B-A3B-FP8",moe_ep_rank="0",pp_rank="0",tp_rank="0"} 2.0
sglang:queue_time_seconds_bucket{engine_type="unified",le="1600.0",model_name="Qwen/Qwen3.5-35B-A3B-FP8",moe_ep_rank="0",pp_rank="0",tp_rank="0"} 2.0
sglang:queue_time_seconds_bucket{engine_type="unified",le="1800.0",model_name="Qwen/Qwen3.5-35B-A3B-FP8",moe_ep_rank="0",pp_rank="0",tp_rank="0"} 2.0
sglang:queue_time_seconds_bucket{engine_type="unified",le="2000.0",model_name="Qwen/Qwen3.5-35B-A3B-FP8",moe_ep_rank="0",pp_rank="0",tp_rank="0"} 2.0
sglang:queue_time_seconds_bucket{engine_type="unified",le="2500.0",model_name="Qwen/Qwen3.5-35B-A3B-FP8",moe_ep_rank="0",pp_rank="0",tp_rank="0"} 2.0
sglang:queue_time_seconds_bucket{engine_type="unified",le="3000.0",model_name="Qwen/Qwen3.5-35B-A3B-FP8",moe_ep_rank="0",pp_rank="0",tp_rank="0"} 2.0
sglang:queue_time_seconds_bucket{engine_type="unified",le="+Inf",model_name="Qwen/Qwen3.5-35B-A3B-FP8",moe_ep_rank="0",pp_rank="0",tp_rank="0"} 2.0
sglang:queue_time_seconds_count{engine_type="unified",model_name="Qwen/Qwen3.5-35B-A3B-FP8",moe_ep_rank="0",pp_rank="0",tp_rank="0"} 2.0
# HELP sglang:cuda_graph_passes_total Total number of forward passes categorized by CUDA graph.
# TYPE sglang:cuda_graph_passes_total counter
sglang:cuda_graph_passes_total{engine_type="unified",mode="prefill_none",model_name="Qwen/Qwen3.5-35B-A3B-FP8",moe_ep_rank="0",pp_rank="0",tp_rank="0"} 2.0
sglang:cuda_graph_passes_total{engine_type="unified",mode="decode_cuda_graph",model_name="Qwen/Qwen3.5-35B-A3B-FP8",moe_ep_rank="0",pp_rank="0",tp_rank="0"} 7.0
# HELP sglang:realtime_tokens_total Total number of tokens processed (updated on each log interval). mode: prefill_compute, prefill_cache, decode.
# TYPE sglang:realtime_tokens_total counter
sglang:realtime_tokens_total{engine_type="unified",mode="prefill_compute",model_name="Qwen/Qwen3.5-35B-A3B-FP8",moe_ep_rank="0",pp_rank="0",tp_rank="0"} 92.0
sglang:realtime_tokens_total{engine_type="unified",mode="decode",model_name="Qwen/Qwen3.5-35B-A3B-FP8",moe_ep_rank="0",pp_rank="0",tp_rank="0"} 7.0
# HELP sglang:time_to_first_token_seconds Histogram of time to first token in seconds.
# TYPE sglang:time_to_first_token_seconds histogram
sglang:time_to_first_token_seconds_sum{model_name="Qwen/Qwen3.5-35B-A3B-FP8"} 35.315967517999994
sglang:time_to_first_token_seconds_bucket{le="0.1",model_name="Qwen/Qwen3.5-35B-A3B-FP8"} 0.0
sglang:time_to_first_token_seconds_bucket{le="0.2",model_name="Qwen/Qwen3.5-35B-A3B-FP8"} 0.0
sglang:time_to_first_token_seconds_bucket{le="0.4",model_name="Qwen/Qwen3.5-35B-A3B-FP8"} 0.0
sglang:time_to_first_token_seconds_bucket{le="0.6",model_name="Qwen/Qwen3.5-35B-A3B-FP8"} 0.0
sglang:time_to_first_token_seconds_bucket{le="0.8",model_name="Qwen/Qwen3.5-35B-A3B-FP8"} 0.0
sglang:time_to_first_token_seconds_bucket{le="1.0",model_name="Qwen/Qwen3.5-35B-A3B-FP8"} 0.0
sglang:time_to_first_token_seconds_bucket{le="2.0",model_name="Qwen/Qwen3.5-35B-A3B-FP8"} 0.0
sglang:time_to_first_token_seconds_bucket{le="4.0",model_name="Qwen/Qwen3.5-35B-A3B-FP8"} 0.0
sglang:time_to_first_token_seconds_bucket{le="6.0",model_name="Qwen/Qwen3.5-35B-A3B-FP8"} 0.0
sglang:time_to_first_token_seconds_bucket{le="8.0",model_name="Qwen/Qwen3.5-35B-A3B-FP8"} 0.0
sglang:time_to_first_token_seconds_bucket{le="10.0",model_name="Qwen/Qwen3.5-35B-A3B-FP8"} 0.0
sglang:time_to_first_token_seconds_bucket{le="20.0",model_name="Qwen/Qwen3.5-35B-A3B-FP8"} 2.0
sglang:time_to_first_token_seconds_bucket{le="40.0",model_name="Qwen/Qwen3.5-35B-A3B-FP8"} 2.0
sglang:time_to_first_token_seconds_bucket{le="60.0",model_name="Qwen/Qwen3.5-35B-A3B-FP8"} 2.0
sglang:time_to_first_token_seconds_bucket{le="80.0",model_name="Qwen/Qwen3.5-35B-A3B-FP8"} 2.0
sglang:time_to_first_token_seconds_bucket{le="100.0",model_name="Qwen/Qwen3.5-35B-A3B-FP8"} 2.0
sglang:time_to_first_token_seconds_bucket{le="200.0",model_name="Qwen/Qwen3.5-35B-A3B-FP8"} 2.0
sglang:time_to_first_token_seconds_bucket{le="400.0",model_name="Qwen/Qwen3.5-35B-A3B-FP8"} 2.0
sglang:time_to_first_token_seconds_bucket{le="+Inf",model_name="Qwen/Qwen3.5-35B-A3B-FP8"} 2.0
sglang:time_to_first_token_seconds_count{model_name="Qwen/Qwen3.5-35B-A3B-FP8"} 2.0
# HELP sglang:e2e_request_latency_seconds Histogram of End-to-end request latency in seconds
# TYPE sglang:e2e_request_latency_seconds histogram
sglang:e2e_request_latency_seconds_sum{model_name="Qwen/Qwen3.5-35B-A3B-FP8"} 35.315979869
sglang:e2e_request_latency_seconds_bucket{le="0.1",model_name="Qwen/Qwen3.5-35B-A3B-FP8"} 0.0
sglang:e2e_request_latency_seconds_bucket{le="0.2",model_name="Qwen/Qwen3.5-35B-A3B-FP8"} 0.0
sglang:e2e_request_latency_seconds_bucket{le="0.4",model_name="Qwen/Qwen3.5-35B-A3B-FP8"} 0.0
sglang:e2e_request_latency_seconds_bucket{le="0.6",model_name="Qwen/Qwen3.5-35B-A3B-FP8"} 0.0
sglang:e2e_request_latency_seconds_bucket{le="0.8",model_name="Qwen/Qwen3.5-35B-A3B-FP8"} 0.0
sglang:e2e_request_latency_seconds_bucket{le="1.0",model_name="Qwen/Qwen3.5-35B-A3B-FP8"} 0.0
sglang:e2e_request_latency_seconds_bucket{le="2.0",model_name="Qwen/Qwen3.5-35B-A3B-FP8"} 0.0
sglang:e2e_request_latency_seconds_bucket{le="4.0",model_name="Qwen/Qwen3.5-35B-A3B-FP8"} 0.0
sglang:e2e_request_latency_seconds_bucket{le="6.0",model_name="Qwen/Qwen3.5-35B-A3B-FP8"} 0.0
sglang:e2e_request_latency_seconds_bucket{le="8.0",model_name="Qwen/Qwen3.5-35B-A3B-FP8"} 0.0
sglang:e2e_request_latency_seconds_bucket{le="10.0",model_name="Qwen/Qwen3.5-35B-A3B-FP8"} 0.0
sglang:e2e_request_latency_seconds_bucket{le="20.0",model_name="Qwen/Qwen3.5-35B-A3B-FP8"} 2.0
sglang:e2e_request_latency_seconds_bucket{le="40.0",model_name="Qwen/Qwen3.5-35B-A3B-FP8"} 2.0
sglang:e2e_request_latency_seconds_bucket{le="60.0",model_name="Qwen/Qwen3.5-35B-A3B-FP8"} 2.0
sglang:e2e_request_latency_seconds_bucket{le="80.0",model_name="Qwen/Qwen3.5-35B-A3B-FP8"} 2.0
sglang:e2e_request_latency_seconds_bucket{le="100.0",model_name="Qwen/Qwen3.5-35B-A3B-FP8"} 2.0
sglang:e2e_request_latency_seconds_bucket{le="200.0",model_name="Qwen/Qwen3.5-35B-A3B-FP8"} 2.0
sglang:e2e_request_latency_seconds_bucket{le="400.0",model_name="Qwen/Qwen3.5-35B-A3B-FP8"} 2.0
sglang:e2e_request_latency_seconds_bucket{le="600.0",model_name="Qwen/Qwen3.5-35B-A3B-FP8"} 2.0
sglang:e2e_request_latency_seconds_bucket{le="1200.0",model_name="Qwen/Qwen3.5-35B-A3B-FP8"} 2.0
sglang:e2e_request_latency_seconds_bucket{le="1800.0",model_name="Qwen/Qwen3.5-35B-A3B-FP8"} 2.0
sglang:e2e_request_latency_seconds_bucket{le="2400.0",model_name="Qwen/Qwen3.5-35B-A3B-FP8"} 2.0
sglang:e2e_request_latency_seconds_bucket{le="+Inf",model_name="Qwen/Qwen3.5-35B-A3B-FP8"} 2.0
sglang:e2e_request_latency_seconds_count{model_name="Qwen/Qwen3.5-35B-A3B-FP8"} 2.0
# HELP sglang:num_retractions Histogram of retraction counts per request.
# TYPE sglang:num_retractions histogram
sglang:num_retractions_sum{model_name="Qwen/Qwen3.5-35B-A3B-FP8"} 0.0
sglang:num_retractions_bucket{le="0.0",model_name="Qwen/Qwen3.5-35B-A3B-FP8"} 2.0
sglang:num_retractions_bucket{le="1.0",model_name="Qwen/Qwen3.5-35B-A3B-FP8"} 2.0
sglang:num_retractions_bucket{le="2.0",model_name="Qwen/Qwen3.5-35B-A3B-FP8"} 2.0
sglang:num_retractions_bucket{le="3.0",model_name="Qwen/Qwen3.5-35B-A3B-FP8"} 2.0
sglang:num_retractions_bucket{le="4.0",model_name="Qwen/Qwen3.5-35B-A3B-FP8"} 2.0
sglang:num_retractions_bucket{le="5.0",model_name="Qwen/Qwen3.5-35B-A3B-FP8"} 2.0
sglang:num_retractions_bucket{le="6.0",model_name="Qwen/Qwen3.5-35B-A3B-FP8"} 2.0
sglang:num_retractions_bucket{le="7.0",model_name="Qwen/Qwen3.5-35B-A3B-FP8"} 2.0
sglang:num_retractions_bucket{le="8.0",model_name="Qwen/Qwen3.5-35B-A3B-FP8"} 2.0
sglang:num_retractions_bucket{le="9.0",model_name="Qwen/Qwen3.5-35B-A3B-FP8"} 2.0
sglang:num_retractions_bucket{le="10.0",model_name="Qwen/Qwen3.5-35B-A3B-FP8"} 2.0
sglang:num_retractions_bucket{le="15.0",model_name="Qwen/Qwen3.5-35B-A3B-FP8"} 2.0
sglang:num_retractions_bucket{le="20.0",model_name="Qwen/Qwen3.5-35B-A3B-FP8"} 2.0
sglang:num_retractions_bucket{le="25.0",model_name="Qwen/Qwen3.5-35B-A3B-FP8"} 2.0
sglang:num_retractions_bucket{le="30.0",model_name="Qwen/Qwen3.5-35B-A3B-FP8"} 2.0
sglang:num_retractions_bucket{le="40.0",model_name="Qwen/Qwen3.5-35B-A3B-FP8"} 2.0
sglang:num_retractions_bucket{le="50.0",model_name="Qwen/Qwen3.5-35B-A3B-FP8"} 2.0
sglang:num_retractions_bucket{le="75.0",model_name="Qwen/Qwen3.5-35B-A3B-FP8"} 2.0
sglang:num_retractions_bucket{le="100.0",model_name="Qwen/Qwen3.5-35B-A3B-FP8"} 2.0
sglang:num_retractions_bucket{le="+Inf",model_name="Qwen/Qwen3.5-35B-A3B-FP8"} 2.0
sglang:num_retractions_count{model_name="Qwen/Qwen3.5-35B-A3B-FP8"} 2.0