added repo
This commit is contained in:
Your Name
@@ -0,0 +1,250 @@
|
||||
# Configuration: Multiple Providers with API Keys and Virtual Keys
|
||||
# Usage: helm install bifrost ./bifrost -f values-examples/providers-and-virtual-keys.yaml
|
||||
#
|
||||
# This example demonstrates:
|
||||
# - Multiple providers (OpenAI, Anthropic, Groq) with 2-3 API keys each
|
||||
# - Virtual keys with provider restrictions
|
||||
# - Budgets and rate limits for governance
|
||||
#
|
||||
# Note: API keys in this example are dummy values for demonstration purposes.
|
||||
# Replace with real keys in production.
|
||||
|
||||
# Image configuration
|
||||
image:
|
||||
repository: docker.io/maximhq/bifrost
|
||||
pullPolicy: IfNotPresent
|
||||
tag: "v1.3.55"
|
||||
|
||||
replicaCount: 1
|
||||
|
||||
# Service
|
||||
service:
|
||||
type: ClusterIP
|
||||
port: 8080
|
||||
|
||||
# Storage configuration - using SQLite for simplicity
|
||||
storage:
|
||||
mode: sqlite
|
||||
persistence:
|
||||
enabled: true
|
||||
size: 5Gi
|
||||
configStore:
|
||||
enabled: true
|
||||
logsStore:
|
||||
enabled: true
|
||||
|
||||
# No PostgreSQL needed for this example
|
||||
postgresql:
|
||||
enabled: false
|
||||
|
||||
# No vector store for this example
|
||||
vectorStore:
|
||||
enabled: false
|
||||
type: none
|
||||
|
||||
# Bifrost configuration
|
||||
bifrost:
|
||||
appDir: /app/data
|
||||
port: 8080
|
||||
host: 0.0.0.0
|
||||
logLevel: info
|
||||
logStyle: json
|
||||
|
||||
client:
|
||||
dropExcessRequests: false
|
||||
initialPoolSize: 100
|
||||
allowedOrigins:
|
||||
- "*"
|
||||
enableLogging: true
|
||||
enforceGovernanceHeader: false
|
||||
allowDirectKeys: false
|
||||
maxRequestBodySizeMb: 100
|
||||
|
||||
# ==========================================================================
|
||||
# PROVIDERS CONFIGURATION
|
||||
# ==========================================================================
|
||||
# Configure multiple providers with 2-3 API keys each.
|
||||
# Keys have weights for load balancing - higher weight = more traffic.
|
||||
# Replace dummy values with real API keys in production.
|
||||
|
||||
providers:
|
||||
# OpenAI - 3 API keys with different weights
|
||||
openai:
|
||||
keys:
|
||||
- name: "openai-primary"
|
||||
value: "sk-dummy-openai-key-1-xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx"
|
||||
weight: 2 # Gets 50% of traffic (2 out of 4 total weight)
|
||||
models:
|
||||
- name: "openai-secondary"
|
||||
value: "sk-dummy-openai-key-2-xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx"
|
||||
weight: 1 # Gets 25% of traffic
|
||||
models:
|
||||
- name: "openai-backup"
|
||||
value: "sk-dummy-openai-key-3-xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx"
|
||||
weight: 1 # Gets 25% of traffic
|
||||
models:
|
||||
|
||||
# Anthropic - 2 API keys
|
||||
anthropic:
|
||||
keys:
|
||||
- name: "anthropic-primary"
|
||||
value: "sk-ant-dummy-key-1-xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx"
|
||||
weight: 1
|
||||
models:
|
||||
- name: "anthropic-secondary"
|
||||
value: "sk-ant-dummy-key-2-xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx"
|
||||
weight: 1
|
||||
models:
|
||||
|
||||
# Groq - 2 API keys
|
||||
groq:
|
||||
keys:
|
||||
- name: "groq-primary"
|
||||
value: "gsk_dummy_groq_key_1_xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx"
|
||||
weight: 1
|
||||
models:
|
||||
- name: "groq-secondary"
|
||||
value: "gsk_dummy_groq_key_2_xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx"
|
||||
weight: 1
|
||||
models:
|
||||
|
||||
# ==========================================================================
|
||||
# GOVERNANCE CONFIGURATION
|
||||
# ==========================================================================
|
||||
# Configure budgets, rate limits, and virtual keys for access control
|
||||
|
||||
governance:
|
||||
# Budget configurations - limit spending per period
|
||||
budgets:
|
||||
- id: "budget-dev"
|
||||
max_limit: 50 # $50 limit
|
||||
reset_duration: "1M" # Resets monthly
|
||||
- id: "budget-production"
|
||||
max_limit: 500 # $500 limit
|
||||
reset_duration: "1M"
|
||||
- id: "budget-testing"
|
||||
max_limit: 10 # $10 limit
|
||||
reset_duration: "1d" # Resets daily
|
||||
|
||||
# Rate limit configurations - limit requests/tokens per period
|
||||
rateLimits:
|
||||
- id: "rate-limit-standard"
|
||||
token_max_limit: 100000
|
||||
token_reset_duration: "1h"
|
||||
request_max_limit: 1000
|
||||
request_reset_duration: "1h"
|
||||
- id: "rate-limit-high"
|
||||
token_max_limit: 500000
|
||||
token_reset_duration: "1h"
|
||||
request_max_limit: 5000
|
||||
request_reset_duration: "1h"
|
||||
- id: "rate-limit-testing"
|
||||
token_max_limit: 10000
|
||||
token_reset_duration: "1h"
|
||||
request_max_limit: 100
|
||||
request_reset_duration: "1h"
|
||||
|
||||
# Virtual Keys - access tokens for different use cases
|
||||
virtualKeys:
|
||||
# Development virtual key - access to ALL providers (no restrictions)
|
||||
- id: "vk-development"
|
||||
name: "Development Key"
|
||||
value: "vk-dev-xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx"
|
||||
is_active: true
|
||||
budget_id: "budget-dev"
|
||||
rate_limit_id: "rate-limit-standard"
|
||||
# No provider_configs means all providers are accessible
|
||||
|
||||
# OpenAI-only virtual key - restricted to OpenAI provider with specific keys
|
||||
- id: "vk-openai-only"
|
||||
name: "OpenAI Only Key"
|
||||
value: "vk-openai-yyyyyyyyyyyyyyyyyyyyyyyyyyyy"
|
||||
is_active: true
|
||||
budget_id: "budget-production"
|
||||
rate_limit_id: "rate-limit-high"
|
||||
provider_configs:
|
||||
- provider: "openai"
|
||||
weight: 1
|
||||
# Restrict to primary and secondary keys only (exclude backup)
|
||||
keys:
|
||||
- name: "openai-primary"
|
||||
- name: "openai-secondary"
|
||||
|
||||
# Anthropic + Groq virtual key - access to both providers with key restrictions
|
||||
- id: "vk-anthropic-groq"
|
||||
name: "Anthropic and Groq Key"
|
||||
value: "vk-anthgroq-zzzzzzzzzzzzzzzzzzzzzzzzzz"
|
||||
is_active: true
|
||||
budget_id: "budget-production"
|
||||
rate_limit_id: "rate-limit-high"
|
||||
provider_configs:
|
||||
- provider: "anthropic"
|
||||
weight: 1
|
||||
# Only use primary anthropic key
|
||||
keys:
|
||||
- name: "anthropic-primary"
|
||||
- provider: "groq"
|
||||
weight: 1
|
||||
# Use both groq keys
|
||||
keys:
|
||||
- name: "groq-primary"
|
||||
- name: "groq-secondary"
|
||||
|
||||
# Testing virtual key - limited budget and rate for testing
|
||||
- id: "vk-testing"
|
||||
name: "Testing Key"
|
||||
value: "vk-test-tttttttttttttttttttttttttttt"
|
||||
is_active: true
|
||||
budget_id: "budget-testing"
|
||||
rate_limit_id: "rate-limit-testing"
|
||||
provider_configs:
|
||||
- provider: "openai"
|
||||
weight: 1
|
||||
allowed_models:
|
||||
- "gpt-4o-mini" # Only allow the cheaper model for testing
|
||||
# Use only the backup key for testing purposes
|
||||
keys:
|
||||
- name: "openai-backup"
|
||||
|
||||
# Plugins configuration
|
||||
plugins:
|
||||
telemetry:
|
||||
enabled: false
|
||||
logging:
|
||||
enabled: true
|
||||
config: {}
|
||||
governance:
|
||||
enabled: true
|
||||
config:
|
||||
is_vk_mandatory: false # Set to true to require virtual key on all requests
|
||||
|
||||
# Resource limits
|
||||
resources:
|
||||
limits:
|
||||
cpu: 1000m
|
||||
memory: 1Gi
|
||||
requests:
|
||||
cpu: 250m
|
||||
memory: 256Mi
|
||||
|
||||
# Probes
|
||||
livenessProbe:
|
||||
httpGet:
|
||||
path: /health
|
||||
port: http
|
||||
initialDelaySeconds: 30
|
||||
periodSeconds: 30
|
||||
timeoutSeconds: 5
|
||||
failureThreshold: 3
|
||||
|
||||
readinessProbe:
|
||||
httpGet:
|
||||
path: /health
|
||||
port: http
|
||||
initialDelaySeconds: 10
|
||||
periodSeconds: 10
|
||||
timeoutSeconds: 5
|
||||
failureThreshold: 3
|
||||
|
||||
autoscaling:
|
||||
enabled: false
|
||||
Reference in New Issue
Block a user