managarten/docker-compose.macmini.yml
Till JS 2a18cb5ee4 feat(mana-ai): v0.7 — cross-tick Deep Research Max pre-planning
Opt-in path for missions that want Gemini Deep Research Max (up to 60 min
per task) instead of the shallow RSS pre-research. Because Max runs well
past a single 60-second tick, the state is carried across ticks:

  tick N:   submit → INSERT mission_research_jobs row → skip planner
  tick N+k: poll → still running → skip planner (metric pending_skips)
  tick N+m: poll → completed → inject as ResolvedInput, DELETE row, plan

- ManaResearchClient talks to mana-research's new internal
  /v1/internal/research/async endpoints with X-Service-Key +
  X-User-Id. Graceful-null on transport errors so a flaky
  mana-research never crashes the tick loop.
- New table mana_ai.mission_research_jobs with PK (user_id, mission_id)
  — presence is the "pending" flag; delete-on-terminal keeps queries
  trivial.
- handleDeepResearch() encapsulates the state machine; planOneMission
  now returns a discriminated union (planned | skipped | failed) so
  "research pending" isn't miscounted as a parse failure.
- Opt-in at TWO gates to keep cost in check ($3–7/task, 1500 credits
  per run):
    1. MANA_AI_DEEP_RESEARCH_ENABLED=true server-side (default off)
    2. DEEP_RESEARCH_TRIGGER regex matches the mission objective
       (strict: "deep research", "tiefe recherche", "umfassende
       recherche", "hintergrundrecherche", "deep dive")
  Falls back to shallow RSS when either gate fails or the submit
  errors upstream.
- Prom metrics: mana_ai_research_jobs_{submitted,completed,failed}_total
  labelled by provider, plus _pending_skips_total.
- docker-compose wires MANA_RESEARCH_URL + the opt-in flag and adds
  mana-research to depends_on.
- Full write-up with real API response shape (outputs plural, not
  OpenAI-style), step-3 MCP-server plan (security-gated, not built),
  ops + kill-switch: docs/reports/gemini-deep-research.md.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-22 17:56:06 +02:00

1754 lines
61 KiB
YAML

# Mana Mac Mini Configuration
# Domain: mana.how (via Cloudflare Tunnel)
#
# Port Schema:
# 3000-3099: Core Services & Backends
# 5000-5099: Web Frontends
# 5100-5199: Games
# 8000-8099: Monitoring Dashboards
# 9000-9199: Infrastructure & Exporters
#
# Naming Convention: mana-{category}-{service}
# Categories: infra, core, app, mon, auto
#
# Memory Limits:
# All containers have explicit mem_limit to prevent unbounded growth.
# Total budget: ~9.8 GiB (fits in 12 GiB Colima VM with ~2 GiB for builds)
# Run ./scripts/mac-mini/memory-baseline.sh to verify actual usage.
# Limits are ceilings — actual usage is typically 50-70% of limits.
services:
# ============================================
# Tier 0: Infrastructure Services
# ============================================
postgres:
image: postgres:16-alpine
container_name: mana-infra-postgres
restart: always
mem_limit: 1024m
environment:
POSTGRES_DB: mana_platform
POSTGRES_USER: postgres
POSTGRES_PASSWORD: ${POSTGRES_PASSWORD:-mana123}
volumes:
- /Volumes/ManaData/postgres:/var/lib/postgresql/data
ports:
- "5432:5432"
healthcheck:
test: ["CMD-SHELL", "pg_isready -U postgres"]
interval: 30s
timeout: 5s
retries: 5
start_period: 10s
command: >
postgres
-c shared_buffers=512MB
-c effective_cache_size=2GB
-c work_mem=16MB
-c maintenance_work_mem=128MB
-c max_wal_size=1GB
-c max_wal_senders=3
-c log_min_duration_statement=1000
-c log_checkpoints=on
# PostgreSQL Backup — hourly pg_dumpall + daily pg_basebackup
# Retention: 48 hourly dumps + 30 daily base backups
# Restore: see docs/POSTGRES_BACKUP.md
postgres-backup:
image: postgres:16-alpine
container_name: mana-infra-postgres-backup
restart: unless-stopped
mem_limit: 128m
depends_on:
postgres:
condition: service_healthy
environment:
PGHOST: postgres
PGUSER: postgres
PGPASSWORD: ${POSTGRES_PASSWORD:-mana123}
BACKUP_DIR: /backups
RETENTION_HOURLY: 48
RETENTION_DAILY: 30
volumes:
- /Volumes/ManaData/backups/postgres:/backups
entrypoint: >
/bin/sh -c "
echo 'PostgreSQL Backup Service started';
echo 'Hourly: pg_dumpall (retention: 48)';
echo 'Daily 03:00: pg_basebackup (retention: 30)';
while true; do
HOUR=$$(date +%H);
TIMESTAMP=$$(date +%Y%m%d_%H%M%S);
echo \"[$$TIMESTAMP] Running hourly backup...\";
pg_dumpall -h postgres -U postgres | gzip > /backups/hourly_$$TIMESTAMP.sql.gz;
if [ \"$$HOUR\" = '03' ]; then
echo \"[$$TIMESTAMP] Running daily base backup...\";
mkdir -p /backups/base_$$TIMESTAMP;
pg_basebackup -h postgres -U postgres -D /backups/base_$$TIMESTAMP -Ft -z -P;
fi;
find /backups -name 'hourly_*.sql.gz' -mmin +$$((48*60)) -delete 2>/dev/null;
find /backups -name 'base_*' -maxdepth 1 -type d -mtime +30 -exec rm -rf {} + 2>/dev/null;
echo \"[$$TIMESTAMP] Backup complete. Sleeping 1h...\";
sleep 3600;
done
"
# Self-hosted Landing Pages (replaces Cloudflare Pages)
# Serves all Astro landing page dist/ folders via Nginx
# Build with: ./scripts/mac-mini/build-landings.sh
landings:
image: nginx:alpine
container_name: mana-infra-landings
restart: always
mem_limit: 48m
volumes:
- ./docker/nginx:/etc/nginx/host-config:ro
- /Volumes/ManaData/landings:/srv/landings:ro
command: >
sh -c "mkdir -p /etc/nginx/snippets &&
cp /etc/nginx/host-config/landings.conf /etc/nginx/conf.d/default.conf &&
cp /etc/nginx/host-config/snippets/* /etc/nginx/snippets/ &&
nginx -g 'daemon off;'"
ports:
- "4400:80"
healthcheck:
test: ["CMD", "wget", "-q", "--spider", "http://127.0.0.1/health"]
interval: 30s
timeout: 5s
retries: 3
redis:
image: redis:7-alpine
container_name: mana-infra-redis
restart: always
mem_limit: 192m
command: redis-server --requirepass ${REDIS_PASSWORD:-redis123}
volumes:
- redis_data:/data
ports:
- "6379:6379"
healthcheck:
test: ["CMD", "redis-cli", "--raw", "incr", "ping"]
interval: 30s
timeout: 5s
retries: 5
start_period: 10s
minio:
image: minio/minio:latest
container_name: mana-infra-minio
restart: always
mem_limit: 256m
command: server /data --console-address ":9001"
environment:
MINIO_ROOT_USER: ${MINIO_ACCESS_KEY:-minioadmin}
MINIO_ROOT_PASSWORD: ${MINIO_SECRET_KEY:-minioadmin}
MINIO_API_CORS_ALLOW_ORIGIN: https://music.mana.how,https://mana.how,https://picture.mana.how,https://storage.mana.how,https://plants.mana.how,https://contacts.mana.how,https://chat.mana.how,https://food.mana.how,https://photos.mana.how
volumes:
- /Volumes/ManaData/minio:/data
ports:
- "9000:9000"
- "9001:9001"
healthcheck:
test: ["CMD", "mc", "ready", "local"]
interval: 30s
timeout: 20s
retries: 3
start_period: 15s
# MinIO bucket initialization and lifecycle rules (runs once)
minio-init:
image: minio/mc:latest
container_name: mana-infra-minio-init
mem_limit: 64m
depends_on:
minio:
condition: service_healthy
entrypoint: >
/bin/sh -c "
mc alias set myminio http://minio:9000 $${MINIO_ACCESS_KEY:-minioadmin} $${MINIO_SECRET_KEY:-minioadmin};
mc mb --ignore-existing myminio/mana-storage;
mc mb --ignore-existing myminio/picture-storage;
mc mb --ignore-existing myminio/chat-storage;
mc mb --ignore-existing myminio/cards-storage;
mc mb --ignore-existing myminio/presi-storage;
mc mb --ignore-existing myminio/calendar-storage;
mc mb --ignore-existing myminio/contacts-storage;
mc mb --ignore-existing myminio/storage-storage;
mc mb --ignore-existing myminio/inventory-storage;
mc mb --ignore-existing myminio/music-storage;
mc mb --ignore-existing myminio/plants-storage;
mc mb --ignore-existing myminio/projectdoc-storage;
mc mb --ignore-existing myminio/mail-storage;
mc anonymous set download myminio/mana-storage;
mc anonymous set download myminio/picture-storage;
mc anonymous set download myminio/plants-storage;
mc anonymous set download myminio/inventory-storage;
mc ilm rule add --expire-days 90 myminio/chat-storage --prefix 'tmp/' 2>/dev/null || true;
mc ilm rule add --expire-days 30 myminio/calendar-storage --prefix 'tmp/' 2>/dev/null || true;
mc ilm rule add --expire-days 7 myminio/picture-storage --prefix 'tmp/' 2>/dev/null || true;
echo 'Buckets and lifecycle rules created successfully';
exit 0;
"
# ============================================
# Tier 0b: Forgejo (Git + CI/CD + Registry)
# ============================================
forgejo:
image: codeberg.org/forgejo/forgejo:11
container_name: mana-core-forgejo
restart: always
mem_limit: 512m
depends_on:
postgres:
condition: service_healthy
environment:
USER_UID: 1000
USER_GID: 1000
FORGEJO__database__DB_TYPE: postgres
FORGEJO__database__HOST: postgres:5432
FORGEJO__database__NAME: forgejo
FORGEJO__database__USER: postgres
FORGEJO__database__PASSWD: ${POSTGRES_PASSWORD:-mana123}
FORGEJO__server__DOMAIN: git.mana.how
FORGEJO__server__SSH_DOMAIN: git.mana.how
FORGEJO__server__ROOT_URL: https://git.mana.how/
FORGEJO__server__HTTP_PORT: 3000
FORGEJO__server__SSH_PORT: 2222
FORGEJO__server__LFS_START_SERVER: "true"
FORGEJO__service__DISABLE_REGISTRATION: "true"
FORGEJO__service__REQUIRE_SIGNIN_VIEW: "false"
FORGEJO__actions__ENABLED: "true"
FORGEJO__actions__DEFAULT_ACTIONS_URL: https://code.forgejo.org
FORGEJO__packages__ENABLED: "true"
FORGEJO__ui__DEFAULT_THEME: forgejo-dark
FORGEJO__ui__SHOW_USER_EMAIL: "false"
FORGEJO__mailer__ENABLED: "false"
volumes:
- /Volumes/ManaData/forgejo:/data
ports:
- "3041:3000"
- "2222:22"
healthcheck:
test: ["CMD", "wget", "-q", "--spider", "http://localhost:3000/api/v1/version"]
interval: 120s
timeout: 10s
retries: 3
start_period: 30s
# Forgejo runner removed — no macOS binary exists, Docker-based runner
# can't access host filesystem/SSH for CD. GitHub CD handles deployment
# via native self-hosted runner. Forgejo is kept as a mirror only.
# ============================================
# Tier 1: Core Auth Service (Port 3001)
# ============================================
mana-auth:
build:
context: .
dockerfile: services/mana-auth/Dockerfile
image: mana-auth:local
container_name: mana-auth
restart: always
mem_limit: 192m
depends_on:
postgres:
condition: service_healthy
environment:
TZ: Europe/Berlin
NODE_ENV: production
PORT: 3001
DATABASE_URL: postgresql://postgres:${POSTGRES_PASSWORD:-mana123}@postgres:5432/mana_platform
BASE_URL: https://auth.mana.how
COOKIE_DOMAIN: .mana.how
MANA_SERVICE_KEY: ${MANA_SERVICE_KEY}
MANA_CREDITS_URL: http://mana-credits:3061
MANA_SUBSCRIPTIONS_URL: http://mana-subscriptions:3063
SYNC_DATABASE_URL: postgresql://postgres:${POSTGRES_PASSWORD:-mana123}@postgres:5432/mana_platform
BETTER_AUTH_SECRET: ${BETTER_AUTH_SECRET:-${JWT_SECRET:-your-jwt-secret-change-me}}
# KEK for the encryption-vault feature (Phase 9). Required in production
# — generate with `openssl rand -base64 32`. See services/mana-auth/CLAUDE.md.
MANA_AUTH_KEK: ${MANA_AUTH_KEK}
# RSA-OAEP-2048 public key of the mana-ai runner. Used to wrap
# per-mission data keys in POST /me/ai-mission-grant. Paired with
# MANA_AI_PRIVATE_KEY_PEM on the mana-ai service. Absent → endpoint
# returns 503 GRANT_NOT_CONFIGURED (graceful degrade).
MANA_AI_PUBLIC_KEY_PEM: ${MANA_AI_PUBLIC_KEY_PEM:-}
MANA_NOTIFY_URL: http://mana-notify:3013
MAX_DAILY_SIGNUPS: ${MAX_DAILY_SIGNUPS:-0}
# Must be a superset of TRUSTED_ORIGINS in
# services/mana-auth/src/auth/better-auth.config.ts.
# Enforced by services/mana-auth/src/auth/sso-config.spec.ts.
# All productivity modules now live under mana.how (path-based) —
# no per-module subdomain entries required here.
CORS_ORIGINS: https://mana.how,https://auth.mana.how,https://arcade.mana.how,https://whopxl.mana.how
ports:
- "3001:3001"
healthcheck:
test: ["CMD", "bun", "-e", "fetch('http://127.0.0.1:3001/health').then(r=>process.exit(r.ok?0:1)).catch(()=>process.exit(1))"]
interval: 120s
timeout: 10s
retries: 3
start_period: 40s
# ============================================
# Tier 1a': AI Mission Runner (Hono + Bun)
# Background ticker that plans due AI Missions and stages proposals
# back to user devices via mana-sync. Opt-in decrypt of encrypted
# inputs via the Mission Key-Grant flow (see services/mana-ai/CLAUDE.md
# and docs/plans/ai-mission-key-grant.md).
# ============================================
mana-ai:
build:
context: .
dockerfile: services/mana-ai/Dockerfile
image: mana-ai:local
container_name: mana-ai
restart: always
mem_limit: 256m
depends_on:
postgres:
condition: service_healthy
mana-llm:
condition: service_started
mana-api:
condition: service_healthy
mana-research:
condition: service_started
environment:
TZ: Europe/Berlin
NODE_ENV: production
PORT: 3067
SYNC_DATABASE_URL: postgresql://postgres:${POSTGRES_PASSWORD:-mana123}@postgres:5432/mana_sync
MANA_LLM_URL: http://mana-llm:3020
MANA_API_URL: http://mana-api:3060
MANA_RESEARCH_URL: http://mana-research:3068
MANA_AI_DEEP_RESEARCH_ENABLED: ${MANA_AI_DEEP_RESEARCH_ENABLED:-false}
MANA_SERVICE_KEY: ${MANA_SERVICE_KEY}
TICK_INTERVAL_MS: ${MANA_AI_TICK_INTERVAL_MS:-60000}
TICK_ENABLED: ${MANA_AI_TICK_ENABLED:-true}
# RSA-OAEP-2048 private key paired with MANA_AI_PUBLIC_KEY_PEM on
# mana-auth. Used to unwrap per-mission data keys at tick time.
# Absent → all grants skip silently with reason="not-configured".
MANA_AI_PRIVATE_KEY_PEM: ${MANA_AI_PRIVATE_KEY_PEM:-}
OTEL_EXPORTER_OTLP_ENDPOINT: http://tempo:4318
ports:
- "3067:3067"
healthcheck:
test: ["CMD", "bun", "-e", "fetch('http://127.0.0.1:3067/health').then(r=>process.exit(r.ok?0:1)).catch(()=>process.exit(1))"]
interval: 120s
timeout: 10s
retries: 3
start_period: 40s
# ============================================
# Tier 1b: Credits Service (Hono + Bun)
# ============================================
mana-credits:
build:
context: .
dockerfile: services/mana-credits/Dockerfile
image: mana-credits:local
container_name: mana-credits
restart: always
mem_limit: 128m
depends_on:
postgres:
condition: service_healthy
environment:
TZ: Europe/Berlin
PORT: 3002
DATABASE_URL: postgresql://postgres:${POSTGRES_PASSWORD:-mana123}@postgres:5432/mana_platform
MANA_AUTH_URL: http://mana-auth:3001
MANA_SERVICE_KEY: ${MANA_SERVICE_KEY}
STRIPE_SECRET_KEY: ${STRIPE_SECRET_KEY:-}
STRIPE_WEBHOOK_SECRET: ${STRIPE_CREDITS_WEBHOOK_SECRET:-}
BASE_URL: https://credits.mana.how
CORS_ORIGINS: https://mana.how,https://chat.mana.how,https://picture.mana.how,https://todo.mana.how,https://quotes.mana.how,https://calendar.mana.how,https://clock.mana.how,https://contacts.mana.how,https://cards.mana.how,https://presi.mana.how,https://storage.mana.how,https://food.mana.how,https://plants.mana.how,https://music.mana.how,https://context.mana.how,https://photos.mana.how,https://questions.mana.how,https://calc.mana.how
ports:
- "3002:3002"
healthcheck:
test: ["CMD", "bun", "-e", "fetch('http://127.0.0.1:3002/health').then(r=>process.exit(r.ok?0:1)).catch(()=>process.exit(1))"]
interval: 120s
timeout: 10s
retries: 3
start_period: 15s
labels:
- "traefik.enable=true"
- "traefik.http.routers.mana-credits.rule=Host(`credits.mana.how`)"
- "traefik.http.routers.mana-credits.tls=true"
- "traefik.http.services.mana-credits.loadbalancer.server.port=3002"
mana-research:
build:
context: .
dockerfile: services/mana-research/Dockerfile
image: mana-research:local
container_name: mana-research
restart: always
mem_limit: 256m
depends_on:
postgres:
condition: service_healthy
redis:
condition: service_started
mana-credits:
condition: service_healthy
mana-search:
condition: service_started
environment:
TZ: Europe/Berlin
NODE_ENV: production
PORT: 3068
DATABASE_URL: postgresql://postgres:${POSTGRES_PASSWORD:-mana123}@postgres:5432/mana_platform
REDIS_URL: redis://redis:6379
MANA_AUTH_URL: http://mana-auth:3001
MANA_LLM_URL: http://mana-llm:3025
MANA_CREDITS_URL: http://mana-credits:3002
MANA_SEARCH_URL: http://mana-search:3021
MANA_SERVICE_KEY: ${MANA_SERVICE_KEY}
CACHE_TTL_SECONDS: 3600
BRAVE_API_KEY: ${BRAVE_API_KEY:-}
TAVILY_API_KEY: ${TAVILY_API_KEY:-}
EXA_API_KEY: ${EXA_API_KEY:-}
SERPER_API_KEY: ${SERPER_API_KEY:-}
JINA_API_KEY: ${JINA_API_KEY:-}
FIRECRAWL_API_KEY: ${FIRECRAWL_API_KEY:-}
SCRAPINGBEE_API_KEY: ${SCRAPINGBEE_API_KEY:-}
PERPLEXITY_API_KEY: ${PERPLEXITY_API_KEY:-}
ANTHROPIC_API_KEY: ${ANTHROPIC_API_KEY:-}
OPENAI_API_KEY: ${OPENAI_API_KEY:-}
GOOGLE_GENAI_API_KEY: ${GOOGLE_GENAI_API_KEY:-}
CORS_ORIGINS: https://mana.how,https://chat.mana.how,https://research.mana.how
ports:
- "3068:3068"
healthcheck:
test: ["CMD", "bun", "-e", "fetch('http://127.0.0.1:3068/health').then(r=>process.exit(r.ok?0:1)).catch(()=>process.exit(1))"]
interval: 120s
timeout: 10s
retries: 3
start_period: 15s
labels:
- "traefik.enable=true"
- "traefik.http.routers.mana-research.rule=Host(`research.mana.how`)"
- "traefik.http.routers.mana-research.tls=true"
- "traefik.http.services.mana-research.loadbalancer.server.port=3068"
mana-events:
build:
context: services/mana-events
dockerfile: Dockerfile
image: mana-events:local
container_name: mana-events
restart: always
mem_limit: 128m
depends_on:
postgres: { condition: service_healthy }
environment:
TZ: Europe/Berlin
PORT: 3065
DATABASE_URL: postgresql://postgres:${POSTGRES_PASSWORD:-mana123}@postgres:5432/mana_platform
MANA_AUTH_URL: http://mana-auth:3001
# The public RSVP endpoints accept requests from anywhere — they're
# designed to be hit by guests who only have the share link.
# The host endpoints sit behind JWT auth so CORS is still scoped.
CORS_ORIGINS: https://mana.how
ports:
- "3065:3065"
healthcheck:
test: ["CMD", "bun", "-e", "fetch('http://127.0.0.1:3065/health').then(r=>process.exit(r.ok?0:1)).catch(()=>process.exit(1))"]
interval: 120s
timeout: 10s
retries: 3
start_period: 15s
labels:
- "traefik.enable=true"
- "traefik.http.routers.mana-events.rule=Host(`events.mana.how`)"
- "traefik.http.routers.mana-events.tls=true"
- "traefik.http.services.mana-events.loadbalancer.server.port=3065"
# ─── Geocoding ───────────────────────────────────────────
# Thin Hono wrapper in front of a self-hosted Pelias stack.
# Pelias itself (elasticsearch + api + libpostal) runs from a separate
# compose file in services/mana-geocoding/pelias/ — see
# services/mana-geocoding/CLAUDE.md for the initial import procedure.
# Internal-only: no traefik labels, not exposed via Cloudflare.
mana-geocoding:
build:
context: services/mana-geocoding
dockerfile: Dockerfile
image: mana-geocoding:local
container_name: mana-geocoding
restart: always
mem_limit: 128m
# Pelias runs on host network via its own compose, so the wrapper
# reaches it via host.docker.internal (Pelias API at :4000).
extra_hosts:
- "host.docker.internal:host-gateway"
environment:
TZ: Europe/Berlin
PORT: 3018
PELIAS_API_URL: http://host.docker.internal:4000/v1
CORS_ORIGINS: https://mana.how,http://localhost:5173
CACHE_MAX_ENTRIES: "5000"
CACHE_TTL_MS: "86400000"
ports:
- "3018:3018"
healthcheck:
test: ["CMD", "bun", "-e", "fetch('http://127.0.0.1:3018/health').then(r=>process.exit(r.ok?0:1)).catch(()=>process.exit(1))"]
interval: 120s
timeout: 10s
retries: 3
start_period: 15s
mana-user:
build:
context: services/mana-user
dockerfile: Dockerfile
image: mana-user:local
container_name: mana-user
restart: always
mem_limit: 128m
depends_on:
postgres: { condition: service_healthy }
environment:
TZ: Europe/Berlin
PORT: 3062
DATABASE_URL: postgresql://postgres:${POSTGRES_PASSWORD:-mana123}@postgres:5432/mana_platform
MANA_AUTH_URL: http://mana-auth:3001
MANA_SERVICE_KEY: ${MANA_SERVICE_KEY}
CORS_ORIGINS: https://mana.how,https://calc.mana.how,https://calendar.mana.how,https://chat.mana.how,https://clock.mana.how,https://contacts.mana.how,https://context.mana.how,https://cards.mana.how,https://music.mana.how,https://food.mana.how,https://photos.mana.how,https://picture.mana.how,https://plants.mana.how,https://presi.mana.how,https://questions.mana.how,https://storage.mana.how,https://todo.mana.how,https://quotes.mana.how
ports:
- "3062:3062"
healthcheck:
test: ["CMD", "bun", "-e", "fetch('http://127.0.0.1:3062/health').then(r=>process.exit(r.ok?0:1)).catch(()=>process.exit(1))"]
interval: 120s
timeout: 10s
retries: 3
start_period: 15s
mana-subscriptions:
build:
context: services/mana-subscriptions
dockerfile: Dockerfile
image: mana-subscriptions:local
container_name: mana-subscriptions
restart: always
mem_limit: 128m
depends_on:
postgres: { condition: service_healthy }
environment:
TZ: Europe/Berlin
PORT: 3063
DATABASE_URL: postgresql://postgres:${POSTGRES_PASSWORD:-mana123}@postgres:5432/mana_platform
MANA_AUTH_URL: http://mana-auth:3001
MANA_SERVICE_KEY: ${MANA_SERVICE_KEY}
STRIPE_SECRET_KEY: ${STRIPE_SECRET_KEY:-}
STRIPE_WEBHOOK_SECRET: ${STRIPE_SUBSCRIPTIONS_WEBHOOK_SECRET:-}
BASE_URL: https://subscriptions.mana.how
CORS_ORIGINS: https://mana.how
ports:
- "3063:3063"
healthcheck:
test: ["CMD", "bun", "-e", "fetch('http://127.0.0.1:3063/health').then(r=>process.exit(r.ok?0:1)).catch(()=>process.exit(1))"]
interval: 120s
timeout: 10s
retries: 3
start_period: 15s
mana-analytics:
build:
context: services/mana-analytics
dockerfile: Dockerfile
image: mana-analytics:local
container_name: mana-analytics
restart: always
mem_limit: 128m
depends_on:
postgres: { condition: service_healthy }
environment:
TZ: Europe/Berlin
PORT: 3064
DATABASE_URL: postgresql://postgres:${POSTGRES_PASSWORD:-mana123}@postgres:5432/mana_platform
MANA_AUTH_URL: http://mana-auth:3001
MANA_LLM_URL: http://mana-llm:3025
CORS_ORIGINS: https://mana.how
ports:
- "3064:3064"
healthcheck:
test: ["CMD", "bun", "-e", "fetch('http://127.0.0.1:3064/health').then(r=>process.exit(r.ok?0:1)).catch(()=>process.exit(1))"]
interval: 120s
timeout: 10s
retries: 3
start_period: 15s
# ============================================
# Tier 2: Gateway & Search Services (Ports 3010-3029)
# ============================================
api-gateway:
build:
context: .
dockerfile: services/mana-api-gateway/Dockerfile
image: mana-api-gateway:local
container_name: mana-api-gateway
restart: always
mem_limit: 64m
depends_on:
postgres:
condition: service_healthy
redis:
condition: service_healthy
environment:
TZ: Europe/Berlin
PORT: 3016
DATABASE_URL: postgresql://postgres:${POSTGRES_PASSWORD:-mana123}@postgres:5432/mana_platform?sslmode=disable
REDIS_HOST: redis
REDIS_PORT: 6379
REDIS_PASSWORD: ${REDIS_PASSWORD:-redis123}
MANA_AUTH_URL: http://mana-auth:3001
SEARCH_SERVICE_URL: http://mana-search:3012
STT_SERVICE_URL: ${STT_SERVICE_URL:-http://192.168.178.11:3020}
TTS_SERVICE_URL: ${TTS_SERVICE_URL:-http://192.168.178.11:3022}
CORS_ORIGINS: https://api.mana.how,https://mana.how
ADMIN_USER_IDS: ${ADMIN_USER_IDS:-}
ports:
- "3016:3016"
healthcheck:
test: ["CMD", "wget", "--no-verbose", "--tries=1", "--spider", "http://127.0.0.1:3016/health"]
interval: 60s
timeout: 5s
retries: 3
start_period: 5s
searxng:
image: searxng/searxng:latest
container_name: mana-core-searxng
restart: always
mem_limit: 256m
volumes:
- ./services/mana-search/searxng:/mnt/searxng-config:ro
entrypoint: ["sh", "-c", "cp /mnt/searxng-config/settings.yml /etc/searxng/settings.yml && cp /mnt/searxng-config/limiter.toml /etc/searxng/limiter.toml 2>/dev/null; exec /usr/local/searxng/entrypoint.sh"]
environment:
SEARXNG_BASE_URL: http://searxng:8080
SEARXNG_SECRET: ${SEARXNG_SECRET:-change-me-searxng-secret}
# Internal only - no external port mapping
healthcheck:
test: ["CMD", "wget", "-q", "--spider", "http://localhost:8080/healthz"]
interval: 120s
timeout: 10s
retries: 3
start_period: 15s
mana-search:
build:
context: .
dockerfile: services/mana-search/Dockerfile
image: mana-search:local
container_name: mana-core-search
restart: always
mem_limit: 64m
depends_on:
searxng:
condition: service_healthy
environment:
PORT: 3012
SEARXNG_URL: http://searxng:8080
SEARXNG_TIMEOUT: 15000
SEARXNG_DEFAULT_LANGUAGE: de-DE
REDIS_HOST: redis
REDIS_PORT: 6379
REDIS_PASSWORD: ${REDIS_PASSWORD:-redis123}
CACHE_SEARCH_TTL: 3600
CACHE_EXTRACT_TTL: 86400
EXTRACT_TIMEOUT: 10000
EXTRACT_MAX_LENGTH: 50000
ports:
- "3012:3012"
healthcheck:
test: ["CMD", "wget", "-q", "--spider", "http://localhost:3012/health"]
interval: 120s
timeout: 10s
retries: 3
start_period: 5s
mana-sync:
build:
context: .
dockerfile: services/mana-sync/Dockerfile
image: mana-sync:local
container_name: mana-core-sync
restart: always
mem_limit: 64m
depends_on:
postgres:
condition: service_healthy
environment:
PORT: 3010
DATABASE_URL: postgresql://postgres:${POSTGRES_PASSWORD:-mana123}@postgres:5432/mana_sync?sslmode=disable
JWKS_URL: http://mana-auth:3001/api/v1/auth/jwks
CORS_ORIGINS: "https://mana.how,https://*.mana.how"
MANA_CREDITS_URL: http://mana-credits:3002
MANA_SERVICE_KEY: ${MANA_SERVICE_KEY}
ports:
- "3010:3010"
healthcheck:
test: ["CMD", "wget", "-q", "--spider", "http://localhost:3010/health"]
interval: 120s
timeout: 5s
retries: 3
start_period: 5s
mana-notify:
build:
context: .
dockerfile: services/mana-notify/Dockerfile
image: mana-notify:local
container_name: mana-core-notify
restart: always
mem_limit: 64m
depends_on:
postgres:
condition: service_healthy
stalwart:
condition: service_started
environment:
PORT: 3013
DATABASE_URL: postgresql://postgres:${POSTGRES_PASSWORD:-mana123}@postgres:5432/mana_platform?sslmode=disable
SERVICE_KEY: ${MANA_SERVICE_KEY:-dev-service-key}
MANA_AUTH_URL: http://mana-auth:3001
SMTP_HOST: ${SMTP_HOST:-stalwart}
SMTP_PORT: ${SMTP_PORT:-587}
SMTP_USER: ${SMTP_USER:-noreply}
SMTP_PASSWORD: ${SMTP_PASSWORD:-ManaNoReply2026!}
SMTP_FROM: "Mana <noreply@mana.how>"
SMTP_INSECURE_TLS: "true"
EXPO_ACCESS_TOKEN: ${EXPO_ACCESS_TOKEN:-}
ports:
- "3013:3013"
healthcheck:
test: ["CMD", "wget", "-q", "--spider", "http://localhost:3013/health"]
interval: 120s
timeout: 5s
retries: 3
start_period: 5s
stalwart:
image: stalwartlabs/stalwart:latest
container_name: mana-mail
restart: always
mem_limit: 256m
ports:
- "25:25"
- "587:587"
- "465:465"
- "993:993"
- "8443:8080"
volumes:
- stalwart_data:/opt/stalwart-mail
environment:
- STALWART_ADMIN_PASSWORD=${STALWART_ADMIN_PASSWORD:-ChangeMe123!}
healthcheck:
test: ["CMD", "wget", "-q", "--spider", "http://localhost:8080/healthz"]
interval: 120s
timeout: 5s
retries: 3
start_period: 10s
mana-crawler:
build:
context: .
dockerfile: services/mana-crawler/Dockerfile
image: mana-crawler:local
container_name: mana-crawler
restart: always
mem_limit: 128m
depends_on:
postgres:
condition: service_healthy
redis:
condition: service_healthy
environment:
TZ: Europe/Berlin
PORT: 3014
DATABASE_URL: postgresql://postgres:${POSTGRES_PASSWORD:-mana123}@postgres:5432/mana_platform?sslmode=disable
REDIS_HOST: redis
REDIS_PORT: 6379
REDIS_PASSWORD: ${REDIS_PASSWORD:-redis123}
CRAWLER_USER_AGENT: "ManaCrawler/1.0 (+https://mana.how/bot)"
QUEUE_CONCURRENCY: 5
ports:
- "3014:3014"
healthcheck:
test: ["CMD", "wget", "--no-verbose", "--tries=1", "--spider", "http://127.0.0.1:3014/health"]
interval: 60s
timeout: 5s
retries: 3
start_period: 5s
mana-media:
build:
context: services/mana-media/apps/api
image: mana-media:local
container_name: mana-core-media
restart: always
mem_limit: 128m
depends_on:
postgres:
condition: service_healthy
redis:
condition: service_healthy
minio:
condition: service_healthy
environment:
NODE_ENV: production
PORT: 3011
DATABASE_URL: postgresql://postgres:${POSTGRES_PASSWORD:-mana123}@postgres:5432/mana_platform
REDIS_HOST: redis
REDIS_PORT: 6379
REDIS_PASSWORD: ${REDIS_PASSWORD:-redis123}
S3_ENDPOINT: minio
S3_PORT: 9000
S3_USE_SSL: "false"
S3_ACCESS_KEY: ${MINIO_ACCESS_KEY:-minioadmin}
S3_SECRET_KEY: ${MINIO_SECRET_KEY:-minioadmin}
S3_BUCKET: mana-media
S3_PUBLIC_URL: https://media.mana.how
PUBLIC_URL: https://media.mana.how/api/v1
CORS_ORIGINS: https://mana.how,https://food.mana.how,https://contacts.mana.how,https://chat.mana.how,https://storage.mana.how,https://photos.mana.how
ports:
- "3011:3011"
healthcheck:
test: ["CMD", "bun", "-e", "fetch('http://127.0.0.1:3011/health').then(r=>process.exit(r.ok?0:1)).catch(()=>process.exit(1))"]
interval: 120s
timeout: 10s
retries: 3
start_period: 10s
mana-landing-builder:
build:
context: .
dockerfile: services/mana-landing-builder/Dockerfile
image: mana-landing-builder:local
container_name: mana-core-landing-builder
restart: always
mem_limit: 192m
depends_on:
mana-auth:
condition: service_healthy
environment:
NODE_ENV: production
PORT: 3015
MANA_AUTH_URL: http://mana-auth:3001
CLOUDFLARE_API_TOKEN: ${CLOUDFLARE_API_TOKEN:-}
CLOUDFLARE_ACCOUNT_ID: ${CLOUDFLARE_ACCOUNT_ID:-}
ORG_LANDING_DOMAIN: mana.how
ports:
- "3015:3015"
healthcheck:
test: ["CMD", "wget", "--no-verbose", "--tries=1", "--spider", "http://127.0.0.1:3015/api/v1/health"]
interval: 120s
timeout: 10s
retries: 3
start_period: 40s
# ============================================
# Tier 3: App Compute Servers (Hono + Bun)
# CRUD is handled by mana-sync. These only handle
# server-side compute (AI, file upload, external APIs).
# ============================================
# --- App Compute Servers (Hono + Bun, ~120 LOC each) ---
# CRUD handled by mana-sync. These handle AI, uploads, external APIs.
# All use FROM oven/bun:1, ~160MB images, ~30MB RAM, <50ms cold start.
# Ports match the old NestJS backends for backward compatibility.
# NOTE: These apps need Hono server Dockerfiles to be added.
# For now they share the same pattern:
# build: { context: apps/{app}/apps/server, dockerfile: ../../Dockerfile.bun }
# Requires: Dockerfile.bun in apps/ root (FROM oven/bun:1, COPY, CMD bun run src/index.ts)
# ============================================
# Tier 5: Web Frontends (Ports 5000-5099)
# ============================================
mana-web:
build:
context: .
dockerfile: apps/mana/apps/web/Dockerfile
args:
PUBLIC_SYNC_SERVER_URL: wss://sync.mana.how
image: mana-web:local
container_name: mana-app-web
restart: always
mem_limit: 256m
depends_on:
mana-auth:
condition: service_healthy
mana-api:
condition: service_healthy
environment:
NODE_ENV: production
PORT: 5000
PUBLIC_MANA_AUTH_URL: http://mana-auth:3001
PUBLIC_MANA_AUTH_URL_CLIENT: https://auth.mana.how
PUBLIC_SYNC_SERVER_URL: http://mana-sync:3010
PUBLIC_SYNC_SERVER_URL_CLIENT: https://sync.mana.how
# Unified Hono/Bun API server (apps/api) — hosts all 16 product
# compute modules (calendar, todo, picture, … who) on port 3060.
# Browser calls go through https://mana-api.mana.how (cloudflared
# tunnel route to mana-api:3060). SSR calls inside the docker
# network use the internal hostname.
#
# NOTE: api.mana.how is already in use for the Go mana-api-gateway
# on port 3016 (different service, predates apps/api). The
# mana-api.* subdomain is the unambiguous new home.
PUBLIC_MANA_API_URL: http://mana-api:3060
PUBLIC_MANA_API_URL_CLIENT: https://mana-api.mana.how
PUBLIC_MANA_CREDITS_URL: http://mana-credits:3002
PUBLIC_MANA_CREDITS_URL_CLIENT: https://credits.mana.how
# Per-app HTTP backend URLs (todo-api, calendar-api, contacts-api,
# chat-api, storage-api, cards-api, music-api, food-api,
# picture-api, presi-api, quotes-api, clock-api, context-api) and
# the standalone memoro-server URL were removed in the pre-launch
# ghost-API cleanup — every product module talks to mana-sync
# directly and the unified `memoro` module is fully local-first.
# See docs/PRE_LAUNCH_CLEANUP.md for the full rationale.
PUBLIC_ULOAD_SERVER_URL: http://uload-server:3070
PUBLIC_ULOAD_SERVER_URL_CLIENT: https://uload-api.mana.how
PUBLIC_MANA_MEDIA_URL: http://mana-media:3011
PUBLIC_MANA_MEDIA_URL_CLIENT: https://media.mana.how
PUBLIC_MANA_LLM_URL: http://mana-llm:3025
PUBLIC_MANA_LLM_URL_CLIENT: https://llm.mana.how
PUBLIC_MANA_EVENTS_URL: http://mana-events:3065
PUBLIC_MANA_EVENTS_URL_CLIENT: https://events.mana.how
# mana-ai background Mission Runner. Browser calls the audit
# endpoint (/api/v1/me/ai-audit) to render the Workbench
# "Datenzugriff" tab. SSR doesn't hit this service directly.
PUBLIC_MANA_AI_URL: http://mana-ai:3067
PUBLIC_MANA_AI_URL_CLIENT: https://mana-ai.mana.how
# Feature flag for the Mission Key-Grant consent UI. false → the
# dialog + Workbench audit tab stay hidden even on missions with
# encrypted inputs. Flip to "true" per deployment once the keypair
# is provisioned (see docs/plans/ai-mission-key-grant.md).
PUBLIC_AI_MISSION_GRANTS: ${PUBLIC_AI_MISSION_GRANTS:-false}
# Analytics & Error Tracking
PUBLIC_UMAMI_WEBSITE_ID: 32777167-e026-4618-933a-3429120b479b
PUBLIC_GLITCHTIP_DSN: ${GLITCHTIP_DSN_MANA_WEB:-}
# Speech-to-Text proxy: SvelteKit /api/v1/voice/transcribe forwards
# to mana-stt via Cloudflare Tunnel. The browser never sees the API
# key — it stays server-side.
MANA_STT_URL: https://gpu-stt.mana.how
MANA_STT_API_KEY: ${MANA_STT_API_KEY:-}
# LLM proxy: /api/v1/voice/parse-task and /api/v1/voice/parse-habit
# call mana-llm for structured extraction. Set WITHOUT the PUBLIC_
# prefix because $env/dynamic/private explicitly excludes vars
# that start with the public prefix — so the parse endpoints
# would never see PUBLIC_MANA_LLM_URL even though it's right
# there in the compose env. Both vars exist; the public one
# is read by the browser-side playground and status page.
MANA_LLM_URL: http://mana-llm:3025
MANA_LLM_API_KEY: ${MANA_LLM_API_KEY:-}
ports:
- "5000:5000"
healthcheck:
test: ["CMD", "wget", "--no-verbose", "--tries=1", "--spider", "http://127.0.0.1:5000/health"]
interval: 180s
timeout: 10s
retries: 3
start_period: 20s
# REMOVED standalone web containers — now served by unified mana-web container (mana.how):
# chat-web, todo-web, quotes-web, calendar-web, clock-web, contacts-web,
# storage-web, presi-web, cards-web, food-web, skilltree-web, photos-web,
# music-web, citycorners-web, picture-web, inventory-web, calc-web, times-web,
# uload-web, memoro-web
# picture-backend: REMOVED — replaced by Hono server (apps/picture/apps/server)
arcade-web:
build:
context: .
dockerfile: games/arcade/apps/web/Dockerfile
image: arcade-web:local
container_name: mana-app-arcade-web
restart: always
mem_limit: 128m
depends_on:
mana-auth:
condition: service_healthy
environment:
NODE_ENV: production
PORT: 5210
PUBLIC_MANA_AUTH_URL: http://mana-auth:3001
PUBLIC_MANA_AUTH_URL_CLIENT: https://auth.mana.how
PUBLIC_SYNC_SERVER_URL: http://mana-sync:3010
PUBLIC_SYNC_SERVER_URL_CLIENT: https://sync.mana.how
ports:
- "5210:5210"
healthcheck:
test: ["CMD", "wget", "--no-verbose", "--tries=1", "--spider", "http://127.0.0.1:5210/health"]
interval: 180s
timeout: 10s
retries: 3
start_period: 20s
manavoxel-web:
build:
context: .
dockerfile: apps/manavoxel/apps/web/Dockerfile
image: manavoxel-web:local
container_name: mana-app-manavoxel-web
restart: always
mem_limit: 128m
depends_on:
mana-auth:
condition: service_healthy
environment:
NODE_ENV: production
PORT: 5028
PUBLIC_MANA_AUTH_URL: http://mana-auth:3001
PUBLIC_MANA_AUTH_URL_CLIENT: https://auth.mana.how
PUBLIC_SYNC_SERVER_URL: http://mana-sync:3010
PUBLIC_SYNC_SERVER_URL_CLIENT: https://sync.mana.how
ports:
- "5028:5028"
healthcheck:
test: ["CMD", "wget", "--no-verbose", "--tries=1", "--spider", "http://127.0.0.1:5028/health"]
interval: 180s
timeout: 10s
retries: 3
start_period: 45s
uload-server:
build:
context: apps/uload/apps/server
dockerfile: Dockerfile
image: uload-server:local
container_name: mana-app-uload-server
restart: always
mem_limit: 256m
depends_on:
postgres:
condition: service_healthy
environment:
NODE_ENV: production
PORT: 3070
DATABASE_URL: postgresql://mana:${POSTGRES_PASSWORD:-devpassword}@postgres:5432/mana_sync
MANA_AUTH_URL: http://mana-auth:3001
CORS_ORIGINS: http://mana-web:5000,https://mana.how,https://uload.mana.how,https://ulo.ad
ports:
- "3070:3070"
healthcheck:
test: ["CMD", "bun", "-e", "fetch('http://127.0.0.1:3070/health').then(r=>process.exit(r.ok?0:1)).catch(()=>process.exit(1))"]
interval: 30s
timeout: 10s
retries: 3
start_period: 10s
memoro-server:
build:
context: .
dockerfile: docker/Dockerfile.hono-server
args:
APP: memoro
image: memoro-server:local
container_name: mana-app-memoro-server
restart: always
mem_limit: 256m
depends_on:
mana-auth:
condition: service_healthy
environment:
NODE_ENV: production
PORT: 3015
MANA_AUTH_URL: http://mana-auth:3001
CORS_ORIGINS: http://mana-web:5000,https://mana.how,https://memoro.mana.how
MEMORO_SUPABASE_URL: ${MEMORO_SUPABASE_URL}
MEMORO_SUPABASE_SERVICE_KEY: ${MEMORO_SUPABASE_SERVICE_KEY}
SERVICE_KEY: ${MEMORO_SERVICE_KEY}
AUDIO_SERVER_URL: http://memoro-audio-server:3016
MANA_LLM_URL: http://mana-llm:3025
MANA_LLM_MODEL: ${MANA_LLM_MODEL:-ollama/gemma3:12b}
GEMINI_API_KEY: ${GEMINI_API_KEY}
AZURE_OPENAI_KEY: ${AZURE_OPENAI_KEY}
AZURE_OPENAI_ENDPOINT: ${AZURE_OPENAI_ENDPOINT}
AZURE_OPENAI_DEPLOYMENT: ${AZURE_OPENAI_DEPLOYMENT}
MANA_CREDITS_URL: http://mana-credits:3006
MANA_CREDITS_SERVICE_KEY: ${MANA_CREDITS_SERVICE_KEY}
ports:
- "3015:3015"
healthcheck:
test: ["CMD", "bun", "-e", "fetch('http://127.0.0.1:3015/health').then(r=>process.exit(r.ok?0:1)).catch(()=>process.exit(1))"]
interval: 30s
timeout: 10s
retries: 3
start_period: 15s
memoro-audio-server:
build:
context: apps/memoro/apps/audio-server
dockerfile: Dockerfile
image: memoro-audio-server:local
container_name: mana-app-memoro-audio-server
restart: always
mem_limit: 512m
environment:
NODE_ENV: production
PORT: 3016
SERVICE_KEY: ${MEMORO_SERVICE_KEY}
MEMORO_SERVER_URL: http://memoro-server:3015
MEMORO_SUPABASE_URL: ${MEMORO_SUPABASE_URL}
MEMORO_SUPABASE_SERVICE_KEY: ${MEMORO_SUPABASE_SERVICE_KEY}
AZURE_SPEECH_KEY_1: ${AZURE_SPEECH_KEY_1}
AZURE_SPEECH_KEY_2: ${AZURE_SPEECH_KEY_2}
AZURE_SPEECH_KEY_3: ${AZURE_SPEECH_KEY_3}
AZURE_SPEECH_KEY_4: ${AZURE_SPEECH_KEY_4}
AZURE_SPEECH_REGION: ${AZURE_SPEECH_REGION:-germanywestcentral}
AZURE_SPEECH_ENDPOINT: ${AZURE_SPEECH_ENDPOINT}
AZURE_STORAGE_ACCOUNT_NAME: ${AZURE_STORAGE_ACCOUNT_NAME}
AZURE_STORAGE_ACCOUNT_KEY: ${AZURE_STORAGE_ACCOUNT_KEY}
AZURE_STORAGE_CONTAINER: ${AZURE_STORAGE_CONTAINER:-memoro-batch-audio}
MANA_STT_URL: http://host.docker.internal:3020
MANA_STT_API_KEY: ${MANA_STT_API_KEY:-}
ports:
- "3016:3016"
healthcheck:
test: ["CMD", "bun", "-e", "fetch('http://127.0.0.1:3016/health').then(r=>process.exit(r.ok?0:1)).catch(()=>process.exit(1))"]
interval: 30s
timeout: 10s
retries: 3
start_period: 20s
mana-llm:
build:
context: ./services/mana-llm
dockerfile: Dockerfile
container_name: mana-service-llm
restart: unless-stopped
mem_limit: 256m
depends_on:
redis:
condition: service_healthy
# Ollama lives on the Windows GPU box at 192.168.178.11:11434, but
# Colima containers can't reach the LAN range — the entire
# 192.168.178.0/24 subnet gets synthesized RST from inside any
# container, even though the macOS host routes there fine. The
# gpu-proxy LaunchAgent on the Mac Mini host (com.mana.gpu-proxy,
# see /Users/mana/gpu-proxy.py) bridges 127.0.0.1:13434 → GPU
# box's 11434, so we go through host.docker.internal:13434 to
# reach Ollama. Without this hop the local mana-llm starts
# cleanly but reports an empty model list and every chat
# completion fails with "All connection attempts failed", which
# cascades into voice quick-add silently degrading to its no-LLM
# fallback for everyone hitting the local stack.
extra_hosts:
- "host.docker.internal:host-gateway"
environment:
PORT: 3025
LOG_LEVEL: info
OLLAMA_URL: ${OLLAMA_URL:-http://host.docker.internal:13434}
OLLAMA_DEFAULT_MODEL: ${OLLAMA_MODEL:-gemma3:12b}
OLLAMA_TIMEOUT: 120
REDIS_URL: redis://redis:6379
OPENROUTER_API_KEY: ${OPENROUTER_API_KEY:-}
GROQ_API_KEY: ${GROQ_API_KEY:-}
TOGETHER_API_KEY: ${TOGETHER_API_KEY:-}
GOOGLE_API_KEY: ${GOOGLE_API_KEY:-}
GOOGLE_DEFAULT_MODEL: gemini-2.0-flash
AUTO_FALLBACK_ENABLED: "true"
OLLAMA_MAX_CONCURRENT: 5
CORS_ORIGINS: https://playground.mana.how,https://mana.how,https://chat.mana.how
ports:
- "3025:3025"
healthcheck:
test: ["CMD", "python", "-c", "import urllib.request; urllib.request.urlopen('http://localhost:3025/health')"]
interval: 120s
timeout: 10s
retries: 3
start_period: 30s
llm-playground:
build:
context: .
dockerfile: apps/playground/apps/web/Dockerfile
container_name: mana-app-llm-playground
restart: unless-stopped
mem_limit: 128m
depends_on:
mana-auth:
condition: service_healthy
mana-llm:
condition: service_healthy
environment:
NODE_ENV: production
PORT: 5050
PUBLIC_MANA_AUTH_URL: http://mana-auth:3001
PUBLIC_MANA_AUTH_URL_CLIENT: https://auth.mana.how
PUBLIC_MANA_LLM_URL: http://mana-llm:3025
PUBLIC_MANA_LLM_URL_CLIENT: https://llm.mana.how
ports:
- "5050:5050"
healthcheck:
test: ["CMD", "wget", "--no-verbose", "--tries=1", "--spider", "http://127.0.0.1:5050/health"]
interval: 180s
timeout: 10s
retries: 3
start_period: 45s
labels:
- "com.centurylinklabs.watchtower.enable=true"
# ============================================
# Tier 7: Monitoring Dashboards (Ports 8000-8099)
# ============================================
grafana:
image: grafana/grafana:10.4.1
container_name: mana-mon-grafana
restart: always
mem_limit: 192m
depends_on:
victoriametrics:
condition: service_healthy
environment:
GF_SECURITY_ADMIN_USER: admin
GF_SECURITY_ADMIN_PASSWORD: ${GRAFANA_PASSWORD:-admin}
GF_USERS_ALLOW_SIGN_UP: false
GF_AUTH_ANONYMOUS_ENABLED: true
GF_AUTH_ANONYMOUS_ORG_ROLE: Viewer
GF_SERVER_ROOT_URL: https://grafana.mana.how
GF_SERVER_HTTP_PORT: 8000
GF_INSTALL_PLUGINS: yesoreyeram-infinity-datasource
GF_DASHBOARDS_DEFAULT_HOME_DASHBOARD_PATH: /var/lib/grafana/dashboards/master-overview.json
volumes:
- ./docker/grafana/provisioning:/etc/grafana/provisioning:ro
- ./docker/grafana/dashboards:/var/lib/grafana/dashboards:ro
- grafana_data:/var/lib/grafana
ports:
- "8000:8000"
healthcheck:
test: ["CMD", "wget", "--no-verbose", "--tries=1", "--spider", "http://127.0.0.1:8000/api/health"]
interval: 300s
timeout: 10s
retries: 3
start_period: 10s
umami:
image: ghcr.io/umami-software/umami:postgresql-latest
container_name: mana-mon-umami
restart: always
mem_limit: 256m
depends_on:
postgres:
condition: service_healthy
environment:
DATABASE_URL: postgresql://postgres:${POSTGRES_PASSWORD:-mana123}@postgres:5432/umami
DATABASE_TYPE: postgresql
APP_SECRET: ${UMAMI_APP_SECRET:-change-me-umami-secret}
DISABLE_TELEMETRY: 1
ports:
- "8010:3000"
healthcheck:
test: ["CMD", "wget", "--no-verbose", "--tries=1", "--spider", "http://127.0.0.1:3000/api/heartbeat"]
interval: 300s
timeout: 10s
retries: 3
start_period: 30s
# ============================================
# Tier 8: Metrics & Exporters (Ports 9000-9199)
# ============================================
victoriametrics:
image: victoriametrics/victoria-metrics:v1.99.0
container_name: mana-mon-victoria
restart: always
mem_limit: 256m
# Mount the host config dir read-only and point promscrape directly at it,
# so edits to docker/prometheus/prometheus.yml are picked up by POST /-/reload
# without a container restart. The previous setup baked a copy into
# /etc/prometheus/ at startup, which silently drifted from the host file
# whenever the container wasn't restarted (matrix removal incident, 2026-04-08).
entrypoint: ["/victoria-metrics-prod", "-storageDataPath=/storage", "-retentionPeriod=2y", "-httpListenAddr=:9090", "-promscrape.config=/etc/prometheus/prometheus.yml", "-promscrape.config.strictParse=false", "-selfScrapeInterval=15s", "-search.latencyOffset=0s"]
volumes:
- ./docker/prometheus:/etc/prometheus:ro
- victoriametrics_data:/storage
ports:
- "9090:9090"
healthcheck:
test: ["CMD", "wget", "--no-verbose", "--tries=1", "--spider", "http://127.0.0.1:9090/health"]
interval: 300s
timeout: 10s
retries: 3
start_period: 10s
tempo:
image: grafana/tempo:2.6.1
container_name: mana-mon-tempo
restart: always
mem_limit: 256m
command: ["-config.file=/etc/tempo/tempo.yaml"]
volumes:
- ./docker/tempo:/etc/tempo:ro
- tempo_data:/var/tempo
ports:
- "4318:4318" # OTLP HTTP receiver
- "3200:3200" # Tempo API (for Grafana)
healthcheck:
test: ["CMD", "wget", "--no-verbose", "--tries=1", "--spider", "http://127.0.0.1:3200/ready"]
interval: 300s
timeout: 10s
retries: 3
start_period: 10s
loki:
image: grafana/loki:3.0.0
container_name: mana-mon-loki
restart: always
mem_limit: 192m
entrypoint: ["sh", "-c", "mkdir -p /etc/loki && cp /mnt/loki-config/*.yaml /etc/loki/ 2>/dev/null; exec /usr/bin/loki -config.file=/etc/loki/local-config.yaml"]
volumes:
- ./docker/loki:/mnt/loki-config:ro
- loki_data:/loki
ports:
- "3100:3100"
healthcheck:
test: ["CMD", "wget", "--no-verbose", "--tries=1", "--spider", "http://127.0.0.1:3100/ready"]
interval: 300s
timeout: 10s
retries: 3
start_period: 15s
promtail:
image: grafana/promtail:3.0.0
container_name: mana-mon-promtail
restart: always
mem_limit: 96m
command: -config.file=/etc/promtail/config.yaml -config.expand-env=true
volumes:
- ./docker/promtail:/etc/promtail:ro
- /var/run/docker.sock:/var/run/docker.sock:ro
depends_on:
loki:
condition: service_started
healthcheck:
test: ["CMD", "wget", "--no-verbose", "--tries=1", "--spider", "http://127.0.0.1:9080/ready"]
interval: 300s
timeout: 10s
retries: 3
start_period: 10s
pushgateway:
image: prom/pushgateway:v1.7.0
container_name: mana-mon-pushgateway
restart: always
mem_limit: 48m
ports:
- "9091:9091"
healthcheck:
test: ["CMD", "wget", "--no-verbose", "--tries=1", "--spider", "http://127.0.0.1:9091/-/healthy"]
interval: 300s
timeout: 10s
retries: 3
start_period: 20s
cadvisor:
image: gcr.io/cadvisor/cadvisor:v0.49.1
container_name: mana-mon-cadvisor
restart: always
mem_limit: 128m
privileged: true
volumes:
- /:/rootfs:ro
- /var/run:/var/run:ro
- /sys:/sys:ro
- /var/lib/docker/:/var/lib/docker:ro
- /dev/disk/:/dev/disk:ro
ports:
- "9110:8080"
healthcheck:
test: ["CMD", "wget", "--no-verbose", "--tries=1", "--spider", "http://127.0.0.1:8080/healthz"]
interval: 300s
timeout: 10s
retries: 3
start_period: 20s
postgres-exporter:
image: prometheuscommunity/postgres-exporter:v0.15.0
container_name: mana-mon-postgres-exporter
restart: always
mem_limit: 48m
depends_on:
postgres:
condition: service_healthy
environment:
DATA_SOURCE_NAME: postgresql://postgres:${POSTGRES_PASSWORD:-mana123}@postgres:5432/postgres?sslmode=disable
ports:
- "9187:9187"
redis-exporter:
image: oliver006/redis_exporter:v1.58.0
container_name: mana-mon-redis-exporter
restart: always
mem_limit: 32m
depends_on:
redis:
condition: service_healthy
environment:
REDIS_ADDR: redis://redis:6379
REDIS_PASSWORD: ${REDIS_PASSWORD:-redis123}
ports:
- "9121:9121"
node-exporter:
image: prom/node-exporter:v1.7.0
container_name: mana-mon-node-exporter
restart: always
mem_limit: 32m
# macOS Docker runs in a Linux VM, so we can only monitor the VM's metrics
# For full host metrics on macOS, install node_exporter natively
command:
- '--collector.disable-defaults'
- '--collector.cpu'
- '--collector.meminfo'
- '--collector.loadavg'
- '--collector.filesystem'
- '--collector.netdev'
- '--collector.time'
- '--collector.uname'
- '--collector.filesystem.mount-points-exclude=^/(sys|proc|dev|host|etc)($$|/)'
ports:
- "9100:9100"
healthcheck:
test: ["CMD", "wget", "--no-verbose", "--tries=1", "--spider", "http://127.0.0.1:9100/metrics"]
interval: 300s
timeout: 10s
retries: 3
start_period: 20s
status-page-gen:
image: alpine:3.20
container_name: mana-status-gen
restart: always
mem_limit: 64m
# host network: direkter Zugriff auf VictoriaMetrics via localhost:9090
# (kein depends_on möglich mit network_mode: host)
network_mode: host
environment:
VICTORIAMETRICS_URL: http://localhost:9090
OUTPUT_FILE: /output/index.html
volumes:
- ./scripts/generate-status-page.sh:/generate.sh:ro
- ./packages/shared-branding/src/mana-apps.ts:/mana-apps.ts:ro
- /Volumes/ManaData/landings/status:/output
command:
- sh
- -c
- |
apk add --no-cache curl jq || { echo "apk add fehlgeschlagen, retry in 10s"; sleep 10; exit 1; }
mkdir -p /output
while true; do
cp /generate.sh /tmp/generate.sh
sh /tmp/generate.sh
sleep 60
done
blackbox-exporter:
image: prom/blackbox-exporter:v0.25.0
container_name: mana-mon-blackbox
restart: always
mem_limit: 128m
# Use Cloudflare + Google public resolvers instead of Docker's
# embedded DNS (127.0.0.11). Docker DNS forwards to the host
# resolver which forwards to the home router (FRITZ!Box), and the
# router keeps a stale negative cache for hours after a hostname
# first fails. New CNAMEs (e.g. fresh GPU public hostnames added
# via the Cloudflare dashboard) appear as "no such host" to the
# blackbox probes for the entire negative-cache TTL even though
# they resolve fine via 1.1.1.1 directly.
dns:
- 1.1.1.1
- 8.8.8.8
command: ["--config.file=/etc/blackbox/blackbox.yml"]
volumes:
- ./docker/blackbox/blackbox.yml:/etc/blackbox/blackbox.yml:ro
ports:
- "9115:9115"
healthcheck:
test: ["CMD", "wget", "--no-verbose", "--tries=1", "--spider", "http://127.0.0.1:9115/"]
interval: 300s
timeout: 10s
retries: 3
start_period: 10s
# ============================================
# Alerting Stack (Ports 9093-9095)
# ============================================
vmalert:
image: victoriametrics/vmalert:v1.99.0
container_name: mana-mon-vmalert
restart: always
mem_limit: 64m
depends_on:
victoriametrics:
condition: service_healthy
alertmanager:
condition: service_healthy
# Same direct-mount pattern as victoriametrics above — see the comment
# there for the rationale.
entrypoint: ["/vmalert-prod", "-datasource.url=http://victoriametrics:9090", "-notifier.url=http://alertmanager:9093", "-remoteWrite.url=http://victoriametrics:9090", "-remoteRead.url=http://victoriametrics:9090", "-rule=/etc/alerts/alerts.yml", "-evaluationInterval=30s", "-httpListenAddr=:8880"]
volumes:
- ./docker/prometheus:/etc/alerts:ro
ports:
- "8880:8880"
healthcheck:
test: ["CMD", "wget", "--no-verbose", "--tries=1", "--spider", "http://127.0.0.1:8880/health"]
interval: 300s
timeout: 10s
retries: 3
start_period: 25s
alertmanager:
image: prom/alertmanager:v0.27.0
container_name: mana-mon-alertmanager
restart: always
mem_limit: 64m
depends_on:
alert-notifier:
condition: service_healthy
command: ["--config.file=/etc/alertmanager/alertmanager.yml", "--storage.path=/alertmanager", "--web.listen-address=:9093"]
volumes:
- ./docker/alertmanager/alertmanager.yml:/etc/alertmanager/alertmanager.yml:ro
- alertmanager_data:/alertmanager
ports:
- "9093:9093"
healthcheck:
test: ["CMD", "wget", "--no-verbose", "--tries=1", "--spider", "http://127.0.0.1:9093/-/healthy"]
interval: 300s
timeout: 10s
retries: 3
start_period: 25s
alert-notifier:
build:
context: ./docker/alert-notifier
dockerfile: Dockerfile
image: alert-notifier:local
container_name: mana-mon-alert-notifier
restart: always
mem_limit: 32m
environment:
PORT: 8080
TELEGRAM_BOT_TOKEN: ${TELEGRAM_BOT_TOKEN:-}
TELEGRAM_CHAT_ID: ${TELEGRAM_CHAT_ID:-}
NTFY_TOPIC: ${NTFY_TOPIC:-}
ports:
- "9095:8080"
healthcheck:
test: ["CMD", "wget", "--no-verbose", "--tries=1", "--spider", "http://127.0.0.1:8080/health"]
interval: 300s
timeout: 5s
retries: 3
start_period: 25s
# ============================================
# Auto-Update (Watchtower)
# ============================================
watchtower:
image: nickfedor/watchtower:latest
container_name: mana-auto-watchtower
restart: always
mem_limit: 64m
volumes:
- /var/run/docker.sock:/var/run/docker.sock
environment:
TZ: Europe/Berlin
DOCKER_API_VERSION: 1.45
WATCHTOWER_POLL_INTERVAL: 300
WATCHTOWER_CLEANUP: "true"
WATCHTOWER_INCLUDE_STOPPED: "false"
# ============================================
# Volumes (Naming: mana-{service}-data)
# ============================================
# ============================================
# GlitchTip Error Tracking (Sentry-compatible)
# ============================================
glitchtip:
image: glitchtip/glitchtip:latest
container_name: mana-mon-glitchtip
restart: always
mem_limit: 256m
environment:
DATABASE_URL: postgres://postgres:${POSTGRES_PASSWORD:-mana123}@postgres:5432/glitchtip
REDIS_URL: redis://:${REDIS_PASSWORD:-redis123}@redis:6379/1
SECRET_KEY: ${GLITCHTIP_SECRET_KEY:-change-me-in-production}
PORT: "8020"
GLITCHTIP_DOMAIN: https://glitchtip.mana.how
DEFAULT_FROM_EMAIL: glitchtip@mana.how
CELERY_WORKER_AUTOSCALE: "1,3"
ENABLE_USER_REGISTRATION: "true"
ports:
- "8020:8020"
depends_on:
postgres:
condition: service_healthy
redis:
condition: service_healthy
healthcheck:
test: ["CMD", "python3", "-c", "import urllib.request; urllib.request.urlopen('http://localhost:8020/_health/')"]
interval: 300s
timeout: 10s
retries: 3
start_period: 30s
glitchtip-worker:
image: glitchtip/glitchtip:latest
container_name: mana-mon-glitchtip-worker
restart: always
mem_limit: 192m
command: ./bin/run-celery-with-beat.sh
environment:
DATABASE_URL: postgres://postgres:${POSTGRES_PASSWORD:-mana123}@postgres:5432/glitchtip
REDIS_URL: redis://:${REDIS_PASSWORD:-redis123}@redis:6379/1
SECRET_KEY: ${GLITCHTIP_SECRET_KEY:-change-me-in-production}
GLITCHTIP_DOMAIN: https://glitchtip.mana.how
CELERY_WORKER_AUTOSCALE: "1,3"
depends_on:
postgres:
condition: service_healthy
redis:
condition: service_healthy
# ============================================
# Unified API Server
# ============================================
# apps/api — Hono/Bun process that hosts all 16 product compute
# modules (calendar, todo, chat, picture, planta, food, news,
# traces, moodlit, presi, music, contacts, storage, context, guides,
# research, who) on a single port. Replaces ~17 per-product backend
# containers from the pre-consolidation era; the unified Mana web
# app's compute calls all flow through here.
mana-api:
build:
context: .
dockerfile: apps/api/Dockerfile
image: mana-api:local
container_name: mana-api
restart: always
mem_limit: 384m
depends_on:
postgres:
condition: service_healthy
mana-auth:
condition: service_healthy
environment:
TZ: Europe/Berlin
NODE_ENV: production
PORT: 3060
# Auth (JWT validation via JWKS)
MANA_AUTH_URL: http://mana-auth:3001
JWT_PUBLIC_KEY: ${JWT_PUBLIC_KEY:-}
# Compute services apps/api orchestrates
MANA_LLM_URL: http://mana-llm:3025
MANA_SEARCH_URL: http://mana-search:3012
MANA_CREDITS_URL: http://mana-credits:3002
MANA_MEDIA_URL: http://mana-media:3011
MANA_CRAWLER_URL: http://mana-crawler:3014
MANA_LLM_DEFAULT_MODEL: ${MANA_LLM_DEFAULT_MODEL:-gemma3:4b}
MANA_SERVICE_KEY: ${MANA_SERVICE_KEY}
# OpenAI — picture module gpt-image-2 path. Optional: without it,
# /api/v1/picture/generate falls through to Replicate/local Flux.
OPENAI_API_KEY: ${OPENAI_API_KEY:-}
# Replicate — fallback for Flux-schnell image generation
REPLICATE_API_TOKEN: ${REPLICATE_API_TOKEN:-}
APP_ID: mana-api
# Database (used by modules that have server-side state — research,
# presi share-links, traces guides). Same Postgres + schema split
# as the rest of the platform.
DATABASE_URL: postgresql://postgres:${POSTGRES_PASSWORD:-mana123}@postgres:5432/mana_platform
# CORS — only the unified mana.how origin needs access today.
# The arcade + manavoxel game frontends don't call apps/api.
CORS_ORIGINS: https://mana.how
# Structured-logger format
LOGGER_FORMAT: json
ports:
- "3060:3060"
healthcheck:
test: ["CMD", "bun", "-e", "fetch('http://127.0.0.1:3060/health').then(r=>process.exit(r.ok?0:1)).catch(()=>process.exit(1))"]
interval: 60s
timeout: 5s
retries: 3
start_period: 30s
# ============================================
# News Ingester
# ============================================
# services/news-ingester — pulls public RSS/JSON feeds into the
# news.curated_articles pool every 15 min. The unified mana-api reads
# from the same table to serve /api/v1/news/feed; user reading lists
# remain client-side in the unified Mana app's local IndexedDB.
news-ingester:
build:
context: services/news-ingester
dockerfile: Dockerfile
image: news-ingester:local
container_name: news-ingester
restart: always
mem_limit: 256m
depends_on:
postgres: { condition: service_healthy }
environment:
TZ: Europe/Berlin
PORT: 3066
DATABASE_URL: postgresql://postgres:${POSTGRES_PASSWORD:-mana123}@postgres:5432/mana_platform
TICK_INTERVAL_MS: "900000"
RUN_ON_STARTUP: "true"
ports:
- "3066:3066"
healthcheck:
test: ["CMD", "bun", "-e", "fetch('http://127.0.0.1:3066/health').then(r=>process.exit(r.ok?0:1)).catch(()=>process.exit(1))"]
interval: 120s
timeout: 10s
retries: 3
start_period: 30s
# ============================================
# Games
# ============================================
# whopixels was removed 2026-04-09 — its core mechanic (LLM-driven
# historical figure guessing) lives now as the `who` module inside
# apps/mana/apps/web. The standalone Phaser/Node container is gone;
# see docs/WHO_MODULE.md for the migration rationale.
volumes:
redis_data:
name: mana-redis-data
victoriametrics_data:
name: mana-victoria-data
alertmanager_data:
name: mana-alertmanager-data
grafana_data:
name: mana-grafana-data
analytics_data:
name: mana-analytics-data
loki_data:
name: mana-loki-data
stalwart_data:
name: mana-stalwart-data
tempo_data:
name: mana-tempo-data