init

docker/chroma/.env.example

@@ -0,0 +1,82 @@
+# =============================================================================
+# Chroma Vector Database Configuration
+# =============================================================================
+# Docs: https://docs.trychroma.com
+# =============================================================================
+
+# ---------------------------------------------------------------------------
+# Image Version
+# ---------------------------------------------------------------------------
+# Pin a specific Chroma version (e.g., 0.6.3). Leave empty or set to 'latest'
+# for the latest release.
+CHROMA_VERSION=latest
+
+# ---------------------------------------------------------------------------
+# Host Port
+# ---------------------------------------------------------------------------
+# The port Chroma's REST API will be accessible on from the host machine.
+# The internal container port is fixed at 8000.
+CHROMA_PORT=8000
+
+# ---------------------------------------------------------------------------
+# Persistence
+# ---------------------------------------------------------------------------
+# Persist data to disk (bind-mounted at /mnt/user/appdata/chroma/data).
+IS_PERSISTENT=TRUE
+PERSIST_DIRECTORY=/chroma/chroma/
+
+# ---------------------------------------------------------------------------
+# Server
+# ---------------------------------------------------------------------------
+# Comma-separated list of origins allowed for CORS (e.g., http://localhost:3000).
+# Leave empty to allow all origins.
+CHROMA_SERVER_CORS_ALLOW_ORIGINS=
+# Thread pool size for the HTTP server.
+CHROMA_SERVER_THREAD_POOL_SIZE=40
+# Allow resetting the database via the API. Set to 'true' with caution — this
+# is destructive and will delete all data.
+ALLOW_RESET=false
+
+# ---------------------------------------------------------------------------
+# Authentication (optional)
+# ---------------------------------------------------------------------------
+# Uncomment and configure to enable authentication. The credentials value
+# depends on the provider — for the basic token provider, use:
+#   CHROMA_SERVER_AUTHN_CREDENTIALS=your-auth-token-here
+#
+# For the multi-user token provider, supply a JSON mapping of tokens to users:
+#   CHROMA_SERVER_AUTHN_CREDENTIALS='{"test-token-1234":"admin-user"}'
+# (sensitive)
+# CHROMA_SERVER_AUTHN_PROVIDER=chromadb.auth.token_authn.TokenAuthenticationServerProvider
+# CHROMA_SERVER_AUTHN_CREDENTIALS=
+# CHROMA_AUTH_TOKEN_TRANSPORT_HEADER=X-Chroma-Token
+
+# ---------------------------------------------------------------------------
+# Authorization (optional)
+# ---------------------------------------------------------------------------
+# Uncomment to enable authorization (requires authentication to be configured).
+# (sensitive)
+# CHROMA_SERVER_AUTHZ_PROVIDER=chromadb.auth.simple_rbac_authz.SimpleRBACAuthorizationProvider
+# CHROMA_SERVER_AUTHZ_CONFIG=
+
+# ---------------------------------------------------------------------------
+# Telemetry
+# ---------------------------------------------------------------------------
+# Disable anonymous telemetry.
+ANONYMIZED_TELEMETRY=false
+# OpenTelemetry endpoint (e.g., http://otel-collector:4318).
+CHROMA_OPEN_TELEMETRY__ENDPOINT=
+CHROMA_OPEN_TELEMETRY__SERVICE_NAME=chromadb
+OTEL_EXPORTER_OTLP_HEADERS=
+
+# ---------------------------------------------------------------------------
+# Migrations
+# ---------------------------------------------------------------------------
+# Migration behavior: 'apply' to run migrations on startup, 'none' to skip.
+MIGRATIONS=apply
+
+# ---------------------------------------------------------------------------
+# Memory / Performance
+# ---------------------------------------------------------------------------
+# Memory limit in bytes for Chroma's internal cache. 0 means unlimited.
+CHROMA_MEMORY_LIMIT_BYTES=0

docker/chroma/compose.yaml

@@ -0,0 +1,94 @@
+name: chroma
+
+services:
+    # ===========================================================================
+    # Chroma Server (vector database)
+    # ===========================================================================
+    # Chroma is the open-source embedding database. It provides a REST API for
+    # storing and querying vector embeddings with metadata filtering, full-text
+    # search, and hybrid search capabilities.
+    #
+    # VoyageAI integration: The Chroma SERVER does not need VOYAGE_API_KEY.
+    # VoyageAI embeddings are configured on the CLIENT side when creating
+    # collections and adding documents. Set VOYAGE_API_KEY in the environment
+    # where client tools run (e.g., the agent host, not this container).
+    #
+    # Docs: https://docs.trychroma.com
+    # ===========================================================================
+    chroma:
+        image: ghcr.io/chroma-core/chroma:${CHROMA_VERSION:-latest}
+        restart: unless-stopped
+        volumes:
+            - ./data:/chroma/chroma/
+        environment:
+            # Persistence — hardcoded to prevent accidental data loss from .env overrides
+            IS_PERSISTENT: TRUE
+            PERSIST_DIRECTORY: /chroma/chroma/
+
+            # Server
+            CHROMA_SERVER_CORS_ALLOW_ORIGINS: ${CHROMA_SERVER_CORS_ALLOW_ORIGINS:-}
+            CHROMA_SERVER_THREAD_POOL_SIZE: ${CHROMA_SERVER_THREAD_POOL_SIZE:-40}
+            CHROMA_SERVER_HTTP_PORT: ${CHROMA_SERVER_HTTP_PORT:-8000}
+            # WARNING: setting ALLOW_RESET=true in .env will wipe all data on restart
+            ALLOW_RESET: false
+
+            # Authentication
+            CHROMA_SERVER_AUTHN_PROVIDER: ${CHROMA_SERVER_AUTHN_PROVIDER:-}
+            CHROMA_SERVER_AUTHN_CREDENTIALS: ${CHROMA_SERVER_AUTHN_CREDENTIALS:-}
+            CHROMA_SERVER_AUTHZ_PROVIDER: ${CHROMA_SERVER_AUTHZ_PROVIDER:-}
+            CHROMA_SERVER_AUTHZ_CONFIG: ${CHROMA_SERVER_AUTHZ_CONFIG:-}
+            CHROMA_AUTH_TOKEN_TRANSPORT_HEADER: ${CHROMA_AUTH_TOKEN_TRANSPORT_HEADER:-}
+
+            # Telemetry
+            ANONYMIZED_TELEMETRY: ${ANONYMIZED_TELEMETRY:-false}
+
+            # Migrations
+            MIGRATIONS: ${MIGRATIONS:-apply}
+
+            # Memory / Performance — sensible default for homelab (~1 GB)
+            # Set to 0 for unlimited (Docker memory limit still applies)
+            CHROMA_MEMORY_LIMIT_BYTES: ${CHROMA_MEMORY_LIMIT_BYTES:-1073741824}
+        ports:
+            - "${CHROMA_PORT:-8000}:8000"
+        # healthcheck:
+        #     test:
+        #         [
+        #             "CMD",
+        #             "python",
+        #             "-c",
+        #             "import sys, urllib.request; r = urllib.request.urlopen('http://localhost:8000/api/v2/heartbeat'); sys.exit(0 if r.status == 200 else 1)",
+        #         ]
+        #     interval: 30s
+        #     timeout: 10s
+        #     retries: 3
+        #     start_period: 15s
+        deploy:
+            resources:
+                limits:
+                    memory: 8G
+                    cpus: "4"
+                reservations:
+                    memory: 2G
+                    cpus: "1"
+        networks:
+            chroma:
+                aliases:
+                    - chroma
+            pipeline:
+                aliases:
+                    - chroma
+
+networks:
+    chroma:
+        name: chroma
+        driver: bridge
+    pipeline:
+        name: pipeline
+        external: true
+    # swag:
+    #     name: swag
+    #     external: true
+
+# Data is persisted via bind mount at ./data/ on the host filesystem.
+# This survives docker compose down -v and container recreation.
+volumes: {}