# HELP litellm_requests_total Total LLM proxy requests
# TYPE litellm_requests_total counter
litellm_requests_total{model="gpt-4-turbo",api_key_alias="prod-primary",status="success"} 482913
litellm_requests_total{model="claude-3-opus",api_key_alias="prod-anthropic",status="success"} 91204
litellm_requests_total{model="llama3",api_key_alias="local-ollama",status="success"} 1203841
# HELP litellm_spend_usd Cumulative spend in USD
# TYPE litellm_spend_usd gauge
litellm_spend_usd{model="gpt-4-turbo",org="sectrace"} 1247.83
litellm_spend_usd{model="claude-3-opus",org="sectrace"} 388.41
# HELP litellm_api_key_info API key metadata
# TYPE litellm_api_key_info gauge
litellm_api_key_info{key_alias="prod-primary",key_prefix="sk-sectr",user="admin@sectrace.org",team="prod"} 1
litellm_api_key_info{key_alias="ci-cd",key_prefix="sk-sectr",user="ci@sectrace.org",team="eng"} 1
# HELP vllm_gpu_cache_usage_perc GPU KV cache usage
# TYPE vllm_gpu_cache_usage_perc gauge
vllm_gpu_cache_usage_perc{model_name="llama3-70b",device="cuda:0"} 0.72
# HELP vllm_num_requests_running Running requests
# TYPE vllm_num_requests_running gauge
vllm_num_requests_running{model_name="llama3-70b"} 3
# HELP llm_proxy_internal_config Internal config (do not scrape in prod)
# TYPE llm_proxy_internal_config gauge
llm_proxy_internal_config{openai_api_key="sk-sectrace-prod-xK9mProd3f8a921c4b7e",db_host="10.0.1.45",redis_host="10.0.1.67"} 1