added repo
This commit is contained in:
Your Name
@@ -0,0 +1,912 @@
|
||||
# Default values for Bifrost
|
||||
# This is a YAML-formatted file.
|
||||
# Declare variables to be passed into your templates.
|
||||
|
||||
# Bifrost application configuration
|
||||
replicaCount: 1
|
||||
|
||||
image:
|
||||
# Container image repository
|
||||
# Default: Docker Hub public image
|
||||
# For enterprise customers with private registry, use full URL:
|
||||
# repository: us-west1-docker.pkg.dev/bifrost-enterprise/your-org/bifrost
|
||||
# repository: your-registry.example.com/your-org/bifrost
|
||||
# repository: 123456789.dkr.ecr.us-east-1.amazonaws.com/bifrost
|
||||
repository: docker.io/maximhq/bifrost
|
||||
pullPolicy: IfNotPresent
|
||||
# REQUIRED: Specify the image tag (e.g., v1.5.0, latest)
|
||||
# Docker images are tagged with v prefix (e.g., v1.5.0)
|
||||
# See available tags at: https://hub.docker.com/r/maximhq/bifrost/tags
|
||||
tag: ""
|
||||
|
||||
imagePullSecrets: []
|
||||
nameOverride: ""
|
||||
fullnameOverride: ""
|
||||
|
||||
serviceAccount:
|
||||
# Specifies whether a service account should be created
|
||||
create: true
|
||||
# Automatically mount a ServiceAccount's API credentials?
|
||||
automount: true
|
||||
# Annotations to add to the service account
|
||||
annotations: {}
|
||||
# The name of the service account to use.
|
||||
# If not set and create is true, a name is generated using the fullname template
|
||||
name: ""
|
||||
|
||||
# Annotations to add to the deployment metadata
|
||||
# Useful for tools like Keel (keel.sh) for automatic image updates
|
||||
# Example:
|
||||
# deploymentAnnotations:
|
||||
# keel.sh/policy: force
|
||||
# keel.sh/trigger: poll
|
||||
deploymentAnnotations: {}
|
||||
|
||||
# Labels to add to the deployment metadata (in addition to default labels)
|
||||
deploymentLabels: {}
|
||||
|
||||
podAnnotations: {}
|
||||
podLabels: {}
|
||||
|
||||
podSecurityContext:
|
||||
fsGroup: 1000
|
||||
runAsUser: 1000
|
||||
runAsNonRoot: true
|
||||
|
||||
securityContext:
|
||||
capabilities:
|
||||
drop:
|
||||
- ALL
|
||||
readOnlyRootFilesystem: false
|
||||
runAsNonRoot: true
|
||||
runAsUser: 1000
|
||||
|
||||
service:
|
||||
type: ClusterIP
|
||||
port: 8080
|
||||
annotations: {}
|
||||
|
||||
ingress:
|
||||
enabled: false
|
||||
className: ""
|
||||
annotations: {}
|
||||
hosts:
|
||||
- host: bifrost.local
|
||||
paths:
|
||||
- path: /
|
||||
pathType: Prefix
|
||||
tls: []
|
||||
|
||||
resources:
|
||||
limits:
|
||||
cpu: 2000m
|
||||
memory: 2Gi
|
||||
requests:
|
||||
cpu: 500m
|
||||
memory: 512Mi
|
||||
|
||||
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
|
||||
minReplicas: 1
|
||||
maxReplicas: 10
|
||||
targetCPUUtilizationPercentage: 80
|
||||
targetMemoryUtilizationPercentage: 80
|
||||
# HPA scaling behavior configuration
|
||||
# Controls how quickly the HPA scales up/down to prevent connection disruption
|
||||
behavior:
|
||||
scaleDown:
|
||||
# Stabilization window prevents rapid scale-down oscillation.
|
||||
# The HPA will wait this long after the last scale event before scaling down again.
|
||||
# Important for long-lived streaming connections (e.g. SSE for LLM inference).
|
||||
stabilizationWindowSeconds: 300
|
||||
policies:
|
||||
- type: Pods
|
||||
value: 1
|
||||
periodSeconds: 120
|
||||
scaleUp:
|
||||
stabilizationWindowSeconds: 30
|
||||
|
||||
# Additional volumes on the output Deployment definition.
|
||||
volumes: []
|
||||
|
||||
# Additional volumeMounts on the output Deployment definition.
|
||||
volumeMounts: []
|
||||
|
||||
nodeSelector: {}
|
||||
|
||||
tolerations: []
|
||||
|
||||
affinity: {}
|
||||
|
||||
# Graceful shutdown configuration for long-lived connections (SSE streaming)
|
||||
# When a pod is terminated (e.g. during HPA scale-down), active streaming connections
|
||||
# are severed abruptly. This causes clients to lose their SSE stream mid-response.
|
||||
# The preStop hook and termination grace period give in-flight requests time to complete.
|
||||
terminationGracePeriodSeconds: 60
|
||||
lifecycle:
|
||||
preStop:
|
||||
exec:
|
||||
# Sleep allows the pod to be removed from the Service endpoints and load balancer
|
||||
# before the process starts shutting down, preventing new connections from arriving
|
||||
# while existing ones drain.
|
||||
command: ["sh", "-c", "sleep 15"]
|
||||
|
||||
# Bifrost specific configuration
|
||||
# You can find entire schema at https://getbifrost.ai/schema
|
||||
bifrost:
|
||||
# Application settings
|
||||
appDir: /app/data
|
||||
port: 8080
|
||||
host: 0.0.0.0
|
||||
logLevel: info
|
||||
logStyle: json
|
||||
|
||||
# Encryption key for sensitive data
|
||||
# Can be set as a secret or environment variable
|
||||
encryptionKey: ""
|
||||
|
||||
# Authentication configuration (top-level)
|
||||
# This controls authentication for Bifrost API and dashboard
|
||||
authConfig:
|
||||
adminUsername: ""
|
||||
adminPassword: ""
|
||||
isEnabled: false
|
||||
disableAuthOnInference: false
|
||||
# Use existing Kubernetes secret for admin credentials
|
||||
existingSecret: ""
|
||||
usernameKey: "username"
|
||||
passwordKey: "password"
|
||||
|
||||
# Client configuration
|
||||
client:
|
||||
dropExcessRequests: false
|
||||
initialPoolSize: 300
|
||||
allowedOrigins:
|
||||
- "*"
|
||||
enableLogging: true
|
||||
disableContentLogging: false
|
||||
disableDbPingsInHealth: false
|
||||
logRetentionDays: 365
|
||||
enforceGovernanceHeader: false
|
||||
allowDirectKeys: false
|
||||
maxRequestBodySizeMb: 100
|
||||
enableLitellmFallbacks: false
|
||||
prometheusLabels: []
|
||||
# Header filtering configuration for x-bf-eh-* headers forwarded to LLM providers
|
||||
headerFilterConfig:
|
||||
allowlist: []
|
||||
denylist: []
|
||||
# asyncJobResultTTL: 3600 # Default TTL for async job results in seconds
|
||||
# requiredHeaders: [] # Headers that must be present on every request
|
||||
# loggingHeaders: [] # Headers to capture in log metadata
|
||||
# allowedHeaders: [] # Additional allowed headers for CORS and WebSocket
|
||||
# mcpAgentDepth: 10 # Maximum depth for MCP agent mode tool execution
|
||||
# mcpToolExecutionTimeout: 30 # Timeout for individual MCP tool execution in seconds
|
||||
# mcpCodeModeBindingLevel: "" # Code mode binding level (server or tool)
|
||||
# mcpToolSyncInterval: 0 # Global tool sync interval in minutes (0 = disabled)
|
||||
# hideDeletedVirtualKeysInFilters: false # Omit deleted virtual keys from logs/MCP filter data
|
||||
|
||||
# Framework configuration
|
||||
framework:
|
||||
pricing:
|
||||
# Custom pricing URL for model cost data
|
||||
pricingUrl: ""
|
||||
# Sync interval in seconds (default: 86400 = 24 hours, minimum: 3600)
|
||||
pricingSyncInterval: 86400
|
||||
|
||||
# Provider configurations (add your provider keys here)
|
||||
# You can specify API keys directly or use env.VAR_NAME syntax to reference environment variables
|
||||
# When using existingSecret in providerSecrets, the keys will be injected as env vars and
|
||||
# you should use env.VAR_NAME syntax in the value field
|
||||
# Note: The entire providers block is passed through to the config as-is.
|
||||
# See https://getbifrost.ai/schema for the full provider schema.
|
||||
providers: {}
|
||||
# openai:
|
||||
# keys:
|
||||
# - name: "primary-key" # Key name (required, must be unique)
|
||||
# value: "sk-..." # Direct value
|
||||
# weight: 1
|
||||
# models: ["gpt-4o", "gpt-4o-mini"] # Restrict key to specific models
|
||||
# use_for_batch_api: false # Whether this key can be used for batch API
|
||||
# - name: "secondary-key"
|
||||
# value: "env.OPENAI_KEY" # Reference to environment variable
|
||||
# weight: 1
|
||||
# # Network configuration (optional, per-provider)
|
||||
# network_config:
|
||||
# base_url: "" # Custom base URL (required for Ollama)
|
||||
# extra_headers: {} # Additional headers to send with requests
|
||||
# default_request_timeout_in_seconds: 300 # Request timeout
|
||||
# max_retries: 3 # Maximum number of retries
|
||||
# retry_backoff_initial_ms: 500 # Initial retry backoff in ms
|
||||
# retry_backoff_max_ms: 5000 # Max retry backoff in ms
|
||||
# # Concurrency configuration (optional)
|
||||
# concurrency_and_buffer_size:
|
||||
# concurrency: 100 # Number of concurrent requests
|
||||
# buffer_size: 200 # Buffer size for requests
|
||||
# # Proxy configuration (optional)
|
||||
# proxy_config:
|
||||
# type: "none" # Options: none, http, socks5, environment
|
||||
# url: ""
|
||||
# username: ""
|
||||
# password: ""
|
||||
# ca_cert_pem: "" # PEM-encoded CA cert for SSL-intercepting proxies
|
||||
# send_back_raw_response: false # Include raw response in BifrostResponse
|
||||
# store_raw_request_response: false # Capture raw payloads for plugins only; not returned to client
|
||||
#
|
||||
# anthropic:
|
||||
# keys:
|
||||
# - name: "anthropic-key"
|
||||
# value: "sk-ant-..."
|
||||
# weight: 1
|
||||
#
|
||||
# # Azure OpenAI example (requires azure_key_config)
|
||||
# azure:
|
||||
# keys:
|
||||
# - name: "azure-key"
|
||||
# value: "..."
|
||||
# weight: 1
|
||||
# azure_key_config:
|
||||
# endpoint: "https://your-resource.openai.azure.com"
|
||||
# api_version: "2024-02-15-preview"
|
||||
# deployments:
|
||||
# gpt-4o: "my-gpt4o-deployment"
|
||||
#
|
||||
# # Google Vertex AI example (requires vertex_key_config)
|
||||
# vertex:
|
||||
# keys:
|
||||
# - name: "vertex-key"
|
||||
# value: ""
|
||||
# weight: 1
|
||||
# vertex_key_config:
|
||||
# project_id: "my-gcp-project"
|
||||
# region: "us-central1"
|
||||
# auth_credentials: "env.GOOGLE_CREDENTIALS"
|
||||
#
|
||||
# # AWS Bedrock example (requires bedrock_key_config)
|
||||
# bedrock:
|
||||
# keys:
|
||||
# - name: "bedrock-key"
|
||||
# value: ""
|
||||
# weight: 1
|
||||
# bedrock_key_config:
|
||||
# region: "us-east-1"
|
||||
# access_key: "env.AWS_ACCESS_KEY_ID"
|
||||
# secret_key: "env.AWS_SECRET_ACCESS_KEY"
|
||||
|
||||
# Provider secrets - use existing Kubernetes secrets for provider API keys
|
||||
# These will be injected as environment variables that can be referenced in providers config
|
||||
providerSecrets: {}
|
||||
# openai:
|
||||
# existingSecret: "my-openai-secret"
|
||||
# key: "api-key"
|
||||
# envVar: "OPENAI_API_KEY" # Environment variable name to inject
|
||||
# anthropic:
|
||||
# existingSecret: "my-anthropic-secret"
|
||||
# key: "api-key"
|
||||
# envVar: "ANTHROPIC_API_KEY"
|
||||
|
||||
# MCP (Model Context Protocol) configuration
|
||||
mcp:
|
||||
enabled: false
|
||||
clientConfigs: []
|
||||
# - name: "example-mcp"
|
||||
# connectionType: "stdio"
|
||||
# stdioConfig:
|
||||
# command: "/path/to/mcp/server"
|
||||
# args: []
|
||||
# envs: []
|
||||
# toolSyncInterval: "10m" # Global tool sync interval (Go duration)
|
||||
# Tool manager configuration
|
||||
toolManagerConfig:
|
||||
toolExecutionTimeout: 30
|
||||
maxAgentDepth: 10
|
||||
# codeModeBindingLevel: "" # Code mode binding level (server or tool)
|
||||
|
||||
# Plugins configuration
|
||||
plugins:
|
||||
telemetry:
|
||||
enabled: false
|
||||
config:
|
||||
custom_labels: []
|
||||
# push_gateway:
|
||||
# enabled: false
|
||||
# push_gateway_url: ""
|
||||
# job_name: "bifrost"
|
||||
# instance_id: ""
|
||||
# push_interval: 15
|
||||
# basic_auth:
|
||||
# username: ""
|
||||
# password: ""
|
||||
|
||||
logging:
|
||||
enabled: false
|
||||
config:
|
||||
disable_content_logging: false
|
||||
logging_headers: []
|
||||
|
||||
governance:
|
||||
enabled: false
|
||||
config:
|
||||
is_vk_mandatory: false
|
||||
required_headers: []
|
||||
is_enterprise: false
|
||||
|
||||
maxim:
|
||||
enabled: false
|
||||
config:
|
||||
api_key: ""
|
||||
log_repo_id: ""
|
||||
# Use existing Kubernetes secret for API key (takes precedence over config.api_key)
|
||||
secretRef:
|
||||
name: ""
|
||||
key: "api-key"
|
||||
|
||||
semanticCache:
|
||||
enabled: false
|
||||
config:
|
||||
# Semantic caching mode (dimension > 1): requires provider, keys, and embedding_model
|
||||
# Direct caching mode (dimension: 1): hash-based exact matching, no embedding provider needed
|
||||
provider: "openai"
|
||||
keys: []
|
||||
embedding_model: "text-embedding-3-small"
|
||||
dimension: 1536
|
||||
threshold: 0.8
|
||||
ttl: "5m"
|
||||
conversation_history_threshold: 3
|
||||
cache_by_model: true
|
||||
cache_by_provider: true
|
||||
exclude_system_prompt: false
|
||||
cleanup_on_shutdown: false
|
||||
vector_store_namespace: ""
|
||||
|
||||
otel:
|
||||
enabled: false
|
||||
config:
|
||||
service_name: "bifrost"
|
||||
collector_url: ""
|
||||
trace_type: "otel"
|
||||
protocol: "grpc"
|
||||
# Push-based metrics export via OTLP (recommended for multi-node clusters)
|
||||
metrics_enabled: false
|
||||
metrics_endpoint: "" # e.g., http://otel-collector:4318/v1/metrics (HTTP) or otel-collector:4317 (gRPC)
|
||||
metrics_push_interval: 15 # Push interval in seconds (1-300)
|
||||
# Custom headers for the collector (supports env.VAR_NAME prefix for env var substitution)
|
||||
headers: {}
|
||||
# TLS configuration
|
||||
tls_ca_cert: "" # Path to TLS CA certificate file
|
||||
insecure: false # Skip TLS verification (ignored if tls_ca_cert is set)
|
||||
|
||||
datadog:
|
||||
enabled: false
|
||||
config:
|
||||
service_name: "bifrost"
|
||||
agent_addr: "localhost:8126"
|
||||
env: ""
|
||||
version: ""
|
||||
custom_tags: {}
|
||||
enable_traces: true
|
||||
|
||||
# Custom/dynamic plugins
|
||||
custom: []
|
||||
# - name: "my-custom-plugin"
|
||||
# enabled: true
|
||||
# path: "/plugins/my-plugin.so"
|
||||
# version: 1
|
||||
# config:
|
||||
# key: value
|
||||
|
||||
# Governance configuration for budgets, rate limits, customers, teams, virtual keys, and routing rules
|
||||
governance:
|
||||
budgets: []
|
||||
# - id: "budget-1"
|
||||
# max_limit: 100
|
||||
# reset_duration: "1M" # Supports: 30s, 5m, 1h, 1d, 1w, 1M, 1Y
|
||||
rateLimits: []
|
||||
# - id: "rate-limit-1"
|
||||
# token_max_limit: 100000
|
||||
# token_reset_duration: "1d"
|
||||
# request_max_limit: 1000
|
||||
# request_reset_duration: "1h"
|
||||
customers: []
|
||||
# - id: "customer-1"
|
||||
# name: "Customer Name"
|
||||
# budget_id: "budget-1"
|
||||
# rate_limit_id: "rate-limit-1"
|
||||
teams: []
|
||||
# - id: "team-1"
|
||||
# name: "Team Name"
|
||||
# customer_id: "customer-1"
|
||||
# budget_id: "budget-1"
|
||||
# rate_limit_id: "rate-limit-1"
|
||||
# profile: {} # Team profile data
|
||||
# config: {} # Team configuration data
|
||||
# claims: {} # Team claims data
|
||||
virtualKeys: []
|
||||
# - id: "vk-1"
|
||||
# name: "Virtual Key 1"
|
||||
# description: "Virtual key description"
|
||||
# value: "vk-..."
|
||||
# is_active: true
|
||||
# team_id: "team-1" # Mutually exclusive with customer_id
|
||||
# customer_id: "" # Mutually exclusive with team_id
|
||||
# budget_id: "budget-1"
|
||||
# rate_limit_id: "rate-limit-1"
|
||||
# # Provider-specific configurations (empty means all providers allowed)
|
||||
# provider_configs:
|
||||
# - provider: "openai"
|
||||
# weight: 1.0
|
||||
# allowed_models: ["gpt-4o"]
|
||||
# budget_id: ""
|
||||
# rate_limit_id: ""
|
||||
# keys:
|
||||
# - key_id: "uuid-of-key"
|
||||
# name: "my-key"
|
||||
# value: "sk-..."
|
||||
# # MCP configurations for this virtual key
|
||||
# mcp_configs:
|
||||
# - mcp_client_id: 1
|
||||
# tools_to_execute: ["tool1", "tool2"]
|
||||
modelConfigs: []
|
||||
# - id: "model-config-1"
|
||||
# model_name: "gpt-4o"
|
||||
# provider: "openai"
|
||||
# budget_id: "budget-1"
|
||||
# rate_limit_id: "rate-limit-1"
|
||||
providers: []
|
||||
# - name: "openai"
|
||||
# budget_id: "budget-1"
|
||||
# rate_limit_id: "rate-limit-1"
|
||||
# send_back_raw_request: false
|
||||
# send_back_raw_response: false
|
||||
routingRules: []
|
||||
# - id: "route-1"
|
||||
# name: "Route to Azure"
|
||||
# description: "Route GPT requests to Azure"
|
||||
# enabled: true
|
||||
# cel_expression: "request.model.startsWith('gpt-')"
|
||||
# provider: "azure"
|
||||
# model: "" # Empty means use original model
|
||||
# fallbacks: ["openai"]
|
||||
# scope: "global" # Options: global, team, customer, virtual_key
|
||||
# scope_id: "" # Required for non-global scopes
|
||||
# priority: 0 # Lower = evaluated first
|
||||
authConfig:
|
||||
adminUsername: ""
|
||||
adminPassword: ""
|
||||
isEnabled: false
|
||||
disableAuthOnInference: false
|
||||
# Use existing Kubernetes secret for admin credentials
|
||||
existingSecret: ""
|
||||
usernameKey: "username"
|
||||
passwordKey: "password"
|
||||
|
||||
# Cluster mode configuration for distributed deployments
|
||||
cluster:
|
||||
enabled: false
|
||||
# region: "" # Region identifier for cluster
|
||||
peers: []
|
||||
# - "bifrost-0.bifrost-headless:7946"
|
||||
# - "bifrost-1.bifrost-headless:7946"
|
||||
gossip:
|
||||
port: 7946
|
||||
config:
|
||||
timeoutSeconds: 10
|
||||
successThreshold: 3
|
||||
failureThreshold: 3
|
||||
discovery:
|
||||
enabled: false
|
||||
# Discovery type: kubernetes, dns, udp, consul, etcd, mdns
|
||||
type: ""
|
||||
allowedAddressSpace: []
|
||||
# Kubernetes discovery
|
||||
k8sNamespace: ""
|
||||
k8sLabelSelector: ""
|
||||
# DNS discovery
|
||||
dnsNames: []
|
||||
# UDP broadcast discovery
|
||||
udpBroadcastPort: 0
|
||||
# Consul discovery
|
||||
consulAddress: ""
|
||||
# Etcd discovery
|
||||
etcdEndpoints: []
|
||||
# mDNS discovery
|
||||
mdnsService: ""
|
||||
|
||||
# SAML/SCIM configuration for enterprise SSO
|
||||
saml:
|
||||
enabled: false
|
||||
# Provider: okta, entra
|
||||
provider: ""
|
||||
config: {}
|
||||
# Okta configuration:
|
||||
# issuerUrl: "https://your-domain.okta.com/oauth2/default"
|
||||
# clientId: ""
|
||||
# clientSecret: ""
|
||||
# audience: ""
|
||||
# userIdField: "sub"
|
||||
# teamIdsField: "groups"
|
||||
# rolesField: "roles"
|
||||
#
|
||||
# Entra (Azure AD) configuration:
|
||||
# tenantId: ""
|
||||
# clientId: ""
|
||||
# clientSecret: ""
|
||||
# cloud: "commercial" # or "gcc-high" or "dod"
|
||||
# audience: ""
|
||||
# appIdUri: ""
|
||||
# userIdField: "oid"
|
||||
# teamIdsField: "groups"
|
||||
# rolesField: "roles"
|
||||
|
||||
# Load balancer configuration for intelligent request routing
|
||||
loadBalancer:
|
||||
enabled: false
|
||||
trackerConfig: {}
|
||||
bootstrap: {}
|
||||
|
||||
# Guardrails configuration for content moderation and policy enforcement
|
||||
guardrails:
|
||||
rules: []
|
||||
# - id: 1
|
||||
# name: "Block PII"
|
||||
# description: "Block requests containing PII"
|
||||
# enabled: true
|
||||
# cel_expression: "!contains(request.body, 'SSN')"
|
||||
# apply_to: "input"
|
||||
# sampling_rate: 100
|
||||
# timeout: 1000
|
||||
providers: []
|
||||
# - id: 1
|
||||
# provider_name: "bedrock"
|
||||
# policy_name: "content-filter"
|
||||
# enabled: true
|
||||
# config: {}
|
||||
|
||||
# Audit logs configuration for CADF-compliant activity logging
|
||||
auditLogs:
|
||||
disabled: false
|
||||
hmacKey: ""
|
||||
|
||||
# Storage configuration
|
||||
storage:
|
||||
# Default storage mode: sqlite or postgres
|
||||
# Used as fallback when per-store type is not specified
|
||||
mode: sqlite # Options: sqlite, postgres
|
||||
|
||||
# Persistent volume for SQLite databases (when using sqlite for any store)
|
||||
persistence:
|
||||
enabled: true
|
||||
# storageClass: "-" # Use default storage class
|
||||
accessMode: ReadWriteOnce
|
||||
size: 10Gi
|
||||
# existingClaim: "" # Use an existing PVC
|
||||
|
||||
# Configuration store settings
|
||||
configStore:
|
||||
enabled: true
|
||||
# Backend type for config store. Empty string uses storage.mode as default
|
||||
type: "" # Options: sqlite, postgres, or "" (uses storage.mode)
|
||||
# PostgreSQL connection pool tuning (only applies when type is postgres)
|
||||
# maxIdleConns: 5
|
||||
# maxOpenConns: 50
|
||||
|
||||
# Logs store settings
|
||||
logsStore:
|
||||
enabled: true
|
||||
# Backend type for logs store. Empty string uses storage.mode as default
|
||||
type: "" # Options: sqlite, postgres, or "" (uses storage.mode)
|
||||
# PostgreSQL connection pool tuning (only applies when type is postgres)
|
||||
# maxIdleConns: 5
|
||||
# maxOpenConns: 50
|
||||
|
||||
# PostgreSQL configuration (when any store uses postgres)
|
||||
postgresql:
|
||||
# Deploy PostgreSQL as part of this chart
|
||||
enabled: false
|
||||
|
||||
# Use external PostgreSQL instance
|
||||
external:
|
||||
enabled: false
|
||||
host: ""
|
||||
port: 5432
|
||||
user: bifrost
|
||||
password: ""
|
||||
database: bifrost
|
||||
sslMode: disable
|
||||
# Use existing Kubernetes secret for password (takes precedence over password field)
|
||||
existingSecret: ""
|
||||
passwordKey: "password"
|
||||
|
||||
# PostgreSQL image configuration
|
||||
image:
|
||||
repository: postgres
|
||||
tag: "16-alpine"
|
||||
pullPolicy: IfNotPresent
|
||||
|
||||
# PostgreSQL subchart configuration (when postgresql.enabled is true)
|
||||
auth:
|
||||
username: bifrost
|
||||
password: bifrost_password
|
||||
database: bifrost
|
||||
|
||||
primary:
|
||||
persistence:
|
||||
enabled: true
|
||||
size: 8Gi
|
||||
|
||||
resources:
|
||||
limits:
|
||||
cpu: 1000m
|
||||
memory: 1Gi
|
||||
requests:
|
||||
cpu: 250m
|
||||
memory: 256Mi
|
||||
|
||||
metrics:
|
||||
enabled: false
|
||||
|
||||
# Vector store configuration
|
||||
vectorStore:
|
||||
# Enable vector store for semantic caching
|
||||
enabled: false
|
||||
type: none # Options: none, weaviate, redis, qdrant
|
||||
|
||||
# Weaviate configuration
|
||||
weaviate:
|
||||
# Deploy Weaviate as part of this chart
|
||||
enabled: false
|
||||
|
||||
# Use external Weaviate instance
|
||||
external:
|
||||
enabled: false
|
||||
scheme: http
|
||||
host: ""
|
||||
apiKey: ""
|
||||
grpcHost: ""
|
||||
grpcSecured: false
|
||||
# timeout: "5s" # Timeout for operations (e.g., "5s", "30s")
|
||||
# className: "" # Class name for vector store
|
||||
# Use existing Kubernetes secret for API key (takes precedence over apiKey field)
|
||||
existingSecret: ""
|
||||
apiKeyKey: "api-key"
|
||||
|
||||
# Weaviate subchart configuration (when weaviate.enabled is true)
|
||||
replicas: 1
|
||||
|
||||
image:
|
||||
repository: semitechnologies/weaviate
|
||||
tag: "1.24.1"
|
||||
|
||||
persistence:
|
||||
enabled: true
|
||||
size: 10Gi
|
||||
|
||||
resources:
|
||||
limits:
|
||||
cpu: 1000m
|
||||
memory: 2Gi
|
||||
requests:
|
||||
cpu: 500m
|
||||
memory: 1Gi
|
||||
|
||||
env:
|
||||
QUERY_DEFAULTS_LIMIT: "25"
|
||||
AUTHENTICATION_ANONYMOUS_ACCESS_ENABLED: "true"
|
||||
PERSISTENCE_DATA_PATH: "/var/lib/weaviate"
|
||||
DEFAULT_VECTORIZER_MODULE: "none"
|
||||
ENABLE_MODULES: ""
|
||||
CLUSTER_HOSTNAME: "node1"
|
||||
|
||||
# Redis configuration
|
||||
redis:
|
||||
# Deploy Redis as part of this chart
|
||||
enabled: false
|
||||
|
||||
# Use external Redis instance
|
||||
external:
|
||||
enabled: false
|
||||
host: ""
|
||||
port: 6379
|
||||
username: ""
|
||||
password: ""
|
||||
database: 0
|
||||
# Connection pool tuning (optional)
|
||||
# poolSize: 10 # Maximum number of socket connections
|
||||
# maxActiveConns: 0 # Maximum number of active connections
|
||||
# minIdleConns: 0 # Minimum number of idle connections
|
||||
# maxIdleConns: 0 # Maximum number of idle connections
|
||||
# connMaxLifetime: "" # Connection max lifetime (e.g., "30m")
|
||||
# connMaxIdleTime: "" # Connection max idle time (e.g., "5m")
|
||||
# dialTimeout: "" # Socket connection timeout (e.g., "5s")
|
||||
# readTimeout: "" # Socket read timeout (e.g., "3s")
|
||||
# writeTimeout: "" # Socket write timeout (e.g., "3s")
|
||||
# contextTimeout: "" # Redis operation timeout (e.g., "10s")
|
||||
# Use existing Kubernetes secret for password (takes precedence over password field)
|
||||
existingSecret: ""
|
||||
passwordKey: "password"
|
||||
|
||||
# Redis image configuration
|
||||
image:
|
||||
repository: redis
|
||||
tag: "7-alpine"
|
||||
pullPolicy: IfNotPresent
|
||||
|
||||
# Redis subchart configuration (when redis.enabled is true)
|
||||
auth:
|
||||
enabled: true
|
||||
password: "redis_password"
|
||||
|
||||
master:
|
||||
persistence:
|
||||
enabled: true
|
||||
size: 8Gi
|
||||
|
||||
resources:
|
||||
limits:
|
||||
cpu: 500m
|
||||
memory: 512Mi
|
||||
requests:
|
||||
cpu: 250m
|
||||
memory: 256Mi
|
||||
|
||||
metrics:
|
||||
enabled: false
|
||||
|
||||
# Qdrant configuration
|
||||
qdrant:
|
||||
# Deploy Qdrant as part of this chart
|
||||
enabled: false
|
||||
|
||||
# Use external Qdrant instance
|
||||
external:
|
||||
enabled: false
|
||||
host: ""
|
||||
port: 6334
|
||||
apiKey: ""
|
||||
useTls: false
|
||||
# Use existing Kubernetes secret for API key (takes precedence over apiKey field)
|
||||
existingSecret: ""
|
||||
apiKeyKey: "api-key"
|
||||
|
||||
# Qdrant image configuration
|
||||
image:
|
||||
repository: qdrant/qdrant
|
||||
tag: "v1.16.0"
|
||||
pullPolicy: IfNotPresent
|
||||
|
||||
# Qdrant subchart configuration (when qdrant.enabled is true)
|
||||
persistence:
|
||||
enabled: true
|
||||
size: 10Gi
|
||||
|
||||
resources:
|
||||
limits:
|
||||
cpu: 1000m
|
||||
memory: 2Gi
|
||||
requests:
|
||||
cpu: 500m
|
||||
memory: 1Gi
|
||||
|
||||
# Pinecone configuration (external only, no self-hosted option)
|
||||
pinecone:
|
||||
external:
|
||||
enabled: false
|
||||
apiKey: ""
|
||||
indexHost: "" # Index host URL from Pinecone console (e.g., your-index.svc.environment.pinecone.io)
|
||||
# Use existing Kubernetes secret for API key (takes precedence over apiKey field)
|
||||
existingSecret: ""
|
||||
apiKeyKey: "api-key"
|
||||
|
||||
# Environment variables
|
||||
env: []
|
||||
# - name: CUSTOM_ENV_VAR
|
||||
# value: "value"
|
||||
|
||||
# Environment variables from secrets/configmaps
|
||||
envFrom: []
|
||||
# - secretRef:
|
||||
# name: my-secret
|
||||
# - configMapRef:
|
||||
# name: my-configmap
|
||||
|
||||
|
||||
# Init containers to run before the main application container.
|
||||
# Provide a list of init containers using standard Kubernetes container spec.
|
||||
initContainers: []
|
||||
|
||||
# --- Meesho Infrastructure Extensions ---
|
||||
|
||||
# PodDisruptionBudget
|
||||
# Ensures high availability during voluntary disruptions (node drains, upgrades)
|
||||
podDisruptionBudget:
|
||||
enabled: false
|
||||
# maxUnavailable and minAvailable are mutually exclusive
|
||||
maxUnavailable: "10%"
|
||||
# minAvailable: ""
|
||||
|
||||
# ExternalSecret (external-secrets.io)
|
||||
# Pulls secrets from Vault and creates a K8s Secret
|
||||
# Use with postgresql.external.existingSecret to inject DB password securely
|
||||
externalSecret:
|
||||
enabled: false
|
||||
# Name of the K8s Secret to create
|
||||
secretName: ""
|
||||
# Vault path to extract secrets from
|
||||
path: ""
|
||||
# Optional: pin to a specific Vault secret version
|
||||
version: ""
|
||||
# Refresh interval ("0" = fetch once, "15s" = poll)
|
||||
refreshInterval: "0"
|
||||
# ClusterSecretStore name
|
||||
secretStoreRef: "vault-backend"
|
||||
|
||||
# Contour HTTPProxy
|
||||
# Replaces standard K8s Ingress with Contour's HTTPProxy for advanced routing
|
||||
httpProxy:
|
||||
enabled: false
|
||||
# Slow start policy applied to all backends
|
||||
slowStart:
|
||||
enabled: false
|
||||
window: "120s"
|
||||
aggression: 1
|
||||
minPercent: 10
|
||||
# List of HTTPProxy resources to create
|
||||
entries: []
|
||||
# - fqdn: "llm-gateway.prd.meesho.int"
|
||||
# ingressClassName: "contour-internal-1"
|
||||
# routes:
|
||||
# - conditions:
|
||||
# - prefix: "/"
|
||||
# services:
|
||||
# - port: 8080
|
||||
# timeoutPolicy:
|
||||
# response: "infinity"
|
||||
|
||||
# KEDA ScaledObject
|
||||
# Replaces standard HPA with KEDA for event-driven autoscaling
|
||||
# IMPORTANT: Set autoscaling.enabled=false when using KEDA (mutually exclusive)
|
||||
keda:
|
||||
enabled: false
|
||||
pollingInterval: 30
|
||||
minReplicaCount: 2
|
||||
maxReplicaCount: 200
|
||||
scaledown:
|
||||
stabilizationWindowSeconds: 1800
|
||||
selectpolicy: Min
|
||||
policies:
|
||||
- type: Pods
|
||||
value: 2
|
||||
periodseconds: 15
|
||||
scaleup:
|
||||
stabilizationWindowSeconds: 120
|
||||
selectpolicy: Max
|
||||
policies:
|
||||
- type: Pods
|
||||
value: 2
|
||||
periodseconds: 15
|
||||
- type: Percent
|
||||
value: 10
|
||||
periodseconds: 15
|
||||
triggers:
|
||||
- type: cpu
|
||||
metricType: Utilization
|
||||
metadata:
|
||||
value: "40"
|
||||
Reference in New Issue
Block a user