fullstack-inferencing/.env.example at main · goabiaryan/fullstack-inferencing · GitHub

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
MODEL_NAME=texttinyllama

# Lambda Cloud + README SSH tunnel: http://127.0.0.1:8000
VLLM_BASE_URL=http://127.0.0.1:8000

VLLM_TLS_VERIFY=true

VLLM_SERVER_PROFILE=baseline

GPU_HOURLY_COST_USD=2.5

LAMBDA_CLOUD_API_KEY=
LAMBDA_INSTANCE_TYPE=
LAMBDA_COST_USE_API=true

# gateway.py bind address (nginx upstream targets this port; add 8766, … for multiple workers)
GATEWAY_HOST=127.0.0.1
GATEWAY_PORT=8765

# Crew / OpenAI client → nginx LB (monitoring/nginx-gateway-lb.conf). Must match listen port there.
GATEWAY_LB_HOST=127.0.0.1
GATEWAY_LB_PORT=8780
# Explicit base wins; if unset, crew defaults to http://GATEWAY_LB_HOST:GATEWAY_LB_PORT/v1
GATEWAY_OPENAI_BASE=http://127.0.0.1:8780/v1
# Set false only if you point crew straight at gateway.py (no nginx)
# GATEWAY_USE_LOAD_BALANCER=false

GATEWAY_METRICS_PORT=9101

OTEL_EXPORTER_OTLP_ENDPOINT=http://127.0.0.1:4317
OTEL_EXPORTER_OTLP_TRACES_ENDPOINT=http://127.0.0.1:4317
OTEL_SERVICE_NAME=gateway

OTEL_TRACES_EXPORTER=none

CREWAI_TRACING_ENABLED=true