HealthLog/docker-compose.yml

# HealthLog — two-container stack (app + Postgres 16).
#
# First run (clean host):
#   cp .env.example .env
#   echo "POSTGRES_PASSWORD=$(openssl rand -hex 32)"   >> .env
#   echo "ENCRYPTION_KEY=$(openssl rand -hex 32)"       >> .env
#   echo "API_TOKEN_HMAC_KEY=$(openssl rand -hex 32)"   >> .env
#   docker compose up -d
#   # then open http://<host>:3000 — the first registered user becomes admin.
#
# LAN / plain-HTTP self-hosts (NAS, homelab, Tailscale): set
#   SESSION_COOKIE_SECURE=false
# in .env, otherwise the session cookie carries the `Secure` flag over HTTP
# and login silently fails. This is the single most common self-host snag.
# Leave it unset (or true) only when a TLS reverse proxy serves HTTPS.
#
# Persistent state lives in exactly one named volume (`pgdata`, Postgres'
# data dir). The app container is stateless, so a backup of the database is
# a backup of everything. The optional GeoLite2 databases are either baked
# into the image at build time or mounted read-only from the host and
# pointed at with GEOLITE2_DIR — either way they are replaceable, not state.

# Shared web/worker environment. `app` and the optional `app-worker` below
# both merge this anchor with `<<: *healthlog_env` instead of each keeping
# their own copy of the list. A hand-copied second list is exactly how the
# app-worker example drifted before: `app` grew to 62 variables over time
# while the worker example stayed at 14, missing every SMTP_*, VAPID_*,
# APNS_*, INSIGHTS_MAX_TOKENS and ALLOW_LOCAL_AI_PRIVATE_HOSTS entry — so a
# deployed worker (the process that sends notifications and runs the AI
# jobs) silently couldn't email, push, or send APNs, and had no local-AI
# host allowlist. One shared block makes that drift structurally
# impossible. HEALTHLOG_PROCESS_TYPE is the one value that's meant to
# differ per service and is overridden after the merge.
x-healthlog-env: &healthlog_env
  # connection_limit is set explicitly: Prisma's own default is
  # cpu_count*2+1, which starves the combined web+worker process
  # (HEALTHLOG_PROCESS_TYPE=all) under background AI/rollup load. Default
  # 20 suits a small host; raise DB_CONNECTION_LIMIT on a busy one (keep it
  # below the database's max_connections, leaving headroom for pg-boss,
  # migrations, and backups). pool_timeout is the seconds a query waits for
  # a free connection before erroring.
  DATABASE_URL: "postgresql://healthlog:${POSTGRES_PASSWORD}@db:5432/healthlog?schema=public&connection_limit=${DB_CONNECTION_LIMIT:-20}&pool_timeout=${DB_POOL_TIMEOUT:-20}"
  NODE_ENV: production
  # Encryption at rest (AES-256-GCM). Generate each secret with
  # `openssl rand -hex 32` (a 64-char hex string = 32 bytes). ENCRYPTION_KEY
  # is the single-key path and is all most self-hosts need. The map form
  # (ENCRYPTION_KEYS={"v1":"<hex>","v2":"<hex>"} + ENCRYPTION_ACTIVE_KEY_ID=v2)
  # is for key rotation — see docs/ops/encryption-key-rotation.md. Keep these
  # values stable and backed up: lose them and the encrypted columns are
  # unrecoverable.
  ENCRYPTION_KEY: "${ENCRYPTION_KEY}"
  ENCRYPTION_KEYS: "${ENCRYPTION_KEYS:-}"
  ENCRYPTION_ACTIVE_KEY_ID: "${ENCRYPTION_ACTIVE_KEY_ID:-}"
  API_TOKEN_HMAC_KEY: "${API_TOKEN_HMAC_KEY}"
  NEXT_PUBLIC_APP_URL: "${NEXT_PUBLIC_APP_URL:-http://localhost:3000}"
  APP_URL: "${APP_URL:-http://localhost:3000}"
  # Session-cookie Secure-flag opt-out. SET THIS TO false FOR LAN / plain-HTTP
  # self-hosts (http://<host>:3000 with no TLS reverse proxy) — otherwise the
  # session cookie carries the `Secure` flag, the browser drops it over HTTP,
  # and login silently fails. Unset → the flag tracks NODE_ENV=production
  # (i.e. on); set to true to force it. Forwarded explicitly so an operator
  # value in .env reaches the Node process (the `environment:` block is a
  # whitelist — vars not listed here don't propagate even though
  # docker-compose reads .env for `${VAR}` substitution at compose-up time).
  SESSION_COOKIE_SECURE: "${SESSION_COOKIE_SECURE:-}"
  NATIVE_CANVAS: "${NATIVE_CANVAS:-}"
  WITHINGS_WEBHOOK_SECRET: "${WITHINGS_WEBHOOK_SECRET:-}"
  # v1.18.11 (W3) — client-IP resolution + geo lookup. Behind Cloudflare
  # the real visitor IP is in `cf-connecting-ip`; set
  # TRUST_CF_CONNECTING_IP=1 ONLY when Cloudflare is in front (off by
  # default — a direct-exposed host must not trust a forgeable header).
  # IP_GEO_LOOKUP_URL overrides the default ipwho.is provider (HTTPS by
  # default); IP_GEO_LOOKUP_DISABLED=1 stops all third-party IP egress.
  # IP_GEO_ALLOW_INSECURE=true permits a plain-HTTP provider URL (e.g. the
  # free ip-api.com endpoint) — off by default, an explicit operator choice.
  # Listed here so operator values in .env reach the container (the
  # `environment:` block is a whitelist — vars not listed never propagate).
  TRUST_CF_CONNECTING_IP: "${TRUST_CF_CONNECTING_IP:-}"
  IP_GEO_LOOKUP_URL: "${IP_GEO_LOOKUP_URL:-}"
  IP_GEO_LOOKUP_DISABLED: "${IP_GEO_LOOKUP_DISABLED:-}"
  IP_GEO_ALLOW_INSECURE: "${IP_GEO_ALLOW_INSECURE:-}"
  # Directory the offline GeoLite2 databases are read from. Default
  # `/opt/geolite2` — the path the published image bakes its own copy
  # into, if the build had a MaxMind licence key. Point this at a
  # directory you mount yourself (see the `volumes:` block below) to
  # bring your own GeoLite2-City.mmdb + GeoLite2-ASN.mmdb, refreshed
  # with `geoipupdate` on your own schedule. Listed here so an operator
  # value in .env reaches the container (the `environment:` block is a
  # whitelist). See docs/self-hosting/geolite2.md.
  GEOLITE2_DIR: "${GEOLITE2_DIR:-/opt/geolite2}"
  # MaxMind licence key for the RUNTIME GeoLite2 fetch (issue #659). Set this
  # and the worker downloads GeoLite2-City + GeoLite2-ASN itself into
  # GEOLITE2_DIR on the next boot and monthly thereafter — no image rebuild,
  # no manual mount. Get a free key at
  # https://www.maxmind.com/en/geolite2/signup (My Account -> Manage License
  # Keys). Leave it unset and nothing is fetched (mounting your own databases
  # or IP_GEO_LOOKUP_DISABLED=1 both still work). Listed here so an operator
  # value in .env reaches the container (the `environment:` block is a
  # whitelist). See docs/self-hosting/geolite2.md.
  MAXMIND_LICENSE_KEY: "${MAXMIND_LICENSE_KEY:-}"
  # all = bundle web + worker in this container (default, single-host)
  # web = serve HTTP only; pair with a separate app-worker service
  HEALTHLOG_PROCESS_TYPE: "${HEALTHLOG_PROCESS_TYPE:-all}"
  BACKUP_ENCRYPTION_KEY: "${BACKUP_ENCRYPTION_KEY:-}"
  BACKUP_S3_ENDPOINT: "${BACKUP_S3_ENDPOINT:-}"
  BACKUP_S3_BUCKET: "${BACKUP_S3_BUCKET:-}"
  BACKUP_S3_ACCESS_KEY: "${BACKUP_S3_ACCESS_KEY:-}"
  BACKUP_S3_SECRET_KEY: "${BACKUP_S3_SECRET_KEY:-}"
  BACKUP_S3_REGION: "${BACKUP_S3_REGION:-auto}"
  BACKUP_RETENTION_DAYS: "${BACKUP_RETENTION_DAYS:-30}"
  # How long the audit trail is kept, in days. Default 365, floor 7 (a value
  # below the floor is ignored rather than obeyed, so a window set in seconds
  # by mistake cannot empty a fresh table). It is not only a retention knob:
  # the shared-access panel PUBLISHES this number — "who entered what stays
  # answerable for N days" is on the dialog that ends somebody's access, and
  # the activity view states the same window. An instance running 90 says 90,
  # so the var has to actually reach the container. Listed here for the reason
  # the block exists: `environment:` is a whitelist, and a value in .env that
  # is not named here never arrives.
  AUDIT_LOG_RETENTION_DAYS: "${AUDIT_LOG_RETENTION_DAYS:-}"
  # Per-export ceiling on MedicationAdministration rows in the FHIR
  # health-record Bundle. Bounded 1..50000, default 5000. Listed here so
  # an operator override in .env actually reaches the container (the
  # `environment:` block is a whitelist).
  FHIR_MAX_MEDICATION_ADMINISTRATIONS: "${FHIR_MAX_MEDICATION_ADMINISTRATIONS:-}"
  # WHOOP integration (optional). The app client id/secret are per-user
  # BYO-keys stored encrypted in the DB, so there is no
  # WHOOP_CLIENT_ID/SECRET env. WHOOP_WEBHOOK_SECRET is the instance-level
  # path-segment + HMAC-verification secret for the webhook;
  # WHOOP_REDIRECT_URI overrides the derived
  # `${NEXT_PUBLIC_APP_URL}/api/whoop/callback` when the public URL
  # differs. Listed here so an operator override in .env reaches the
  # container (the `environment:` block is a whitelist).
  WHOOP_WEBHOOK_SECRET: "${WHOOP_WEBHOOK_SECRET:-}"
  WHOOP_REDIRECT_URI: "${WHOOP_REDIRECT_URI:-}"
  # v1.17.0 (F4) — Polar AccessLink + Oura Cloud OAuth (shared app from
  # env). Both pairs together enable each connect button; an unset pair
  # leaves that integration disabled. The optional *_REDIRECT_URI overrides
  # the derived `${NEXT_PUBLIC_APP_URL}/api/{polar,oura}/callback`. Listed
  # here so operator values in .env reach the container (the `environment:`
  # block is a whitelist).
  POLAR_CLIENT_ID: "${POLAR_CLIENT_ID:-}"
  POLAR_CLIENT_SECRET: "${POLAR_CLIENT_SECRET:-}"
  POLAR_REDIRECT_URI: "${POLAR_REDIRECT_URI:-}"
  OURA_CLIENT_ID: "${OURA_CLIENT_ID:-}"
  OURA_CLIENT_SECRET: "${OURA_CLIENT_SECRET:-}"
  OURA_REDIRECT_URI: "${OURA_REDIRECT_URI:-}"
  # v1.28.x — Strava OAuth (shared app from env; per-user BYO keys override).
  # Both together enable the connect button; an unset pair leaves it
  # disabled. STRAVA_REDIRECT_URI overrides the derived
  # `${NEXT_PUBLIC_APP_URL}/api/strava/callback`. Listed here so operator
  # values in .env reach the container (the `environment:` block is a
  # whitelist).
  STRAVA_CLIENT_ID: "${STRAVA_CLIENT_ID:-}"
  STRAVA_CLIENT_SECRET: "${STRAVA_CLIENT_SECRET:-}"
  STRAVA_REDIRECT_URI: "${STRAVA_REDIRECT_URI:-}"
  # OIDC SSO login (self-hosted IdP: Authentik/Keycloak/Authelia/etc).
  # All three of ISSUER_URL/CLIENT_ID/CLIENT_SECRET together enable the
  # "Sign in with SSO" button; set none to leave it disabled. OIDC_ONLY
  # hides password/passkey login entirely, but only takes effect when
  # the three above are also fully set — a half-configured group never
  # locks every user out. It IS enforced server-side, and there is no
  # OIDC flow for the native iOS client yet — enabling it locks iOS
  # users out of sign-in entirely. Listed here so operator values in
  # .env reach the container (the `environment:` block is a whitelist).
  OIDC_ISSUER_URL: "${OIDC_ISSUER_URL:-}"
  OIDC_CLIENT_ID: "${OIDC_CLIENT_ID:-}"
  OIDC_CLIENT_SECRET: "${OIDC_CLIENT_SECRET:-}"
  OIDC_SCOPES: "${OIDC_SCOPES:-}"
  OIDC_BUTTON_LABEL: "${OIDC_BUTTON_LABEL:-}"
  OIDC_ONLY: "${OIDC_ONLY:-}"
  # Deploy-result webhook (optional). DEPLOY_WEBHOOK_SECRET authenticates
  # POST /api/internal/deploy-webhook; DEPLOY_LOGS_URL is the
  # deployment-dashboard URL (e.g. the Coolify UI) linked in
  # deploy-failure notifications. Listed here so operator values in
  # .env reach the container (the `environment:` block is a whitelist).
  DEPLOY_WEBHOOK_SECRET: "${DEPLOY_WEBHOOK_SECRET:-}"
  DEPLOY_LOGS_URL: "${DEPLOY_LOGS_URL:-}"
  # Optional replay hardening: when "true", the deploy webhook
  # requires a fresh X-Deploy-Webhook-Timestamp header (±5 min).
  # See docs/ops/deploy.md.
  DEPLOY_WEBHOOK_REQUIRE_TIMESTAMP: "${DEPLOY_WEBHOOK_REQUIRE_TIMESTAMP:-}"
  # TLS leaf SPKI-change alarm baseline (optional). The native client
  # pins the served TLS leaf; this is the known-good pin set the monitor
  # alarms against on a rotation. Comma-separated for the dual-pin
  # renewal window. Listed here so an operator value in .env reaches the
  # container (the `environment:` block is a whitelist). See
  # docs/ops/tls-cert-pin.md.
  TLS_LEAF_SPKI_PINS: "${TLS_LEAF_SPKI_PINS:-}"
  # Local-AI SSRF opt-in (optional). The local AI provider rejects
  # private/internal hosts by default; set this to an exact host allowlist
  # (e.g. "ollama.lan,10.0.0.5") or "true" for any private host, intended
  # for a self-hosted Ollama / LM Studio. Listed here so an operator value
  # in .env reaches the container (the `environment:` block is a whitelist).
  # See docs/integrations/ai-providers.md.
  ALLOW_LOCAL_AI_PRIVATE_HOSTS: "${ALLOW_LOCAL_AI_PRIVATE_HOSTS:-}"
  # Daily-briefing output-token ceiling (optional; default 2500, clamped
  # 500-8000). Raise when a verbose model's briefing JSON is cut off.
  # Listed here so an operator value in .env reaches the container (the
  # `environment:` block is a whitelist). See docs/integrations/ai-providers.md.
  INSIGHTS_MAX_TOKENS: "${INSIGHTS_MAX_TOKENS:-}"
  # Per-user hourly ceiling for POST /api/insights/generate (optional;
  # default 10). Was documented but missing from this whitelist, so an
  # operator value never reached the container. See
  # docs/integrations/ai-providers.md.
  INSIGHTS_RATE_LIMIT_PER_HOUR: "${INSIGHTS_RATE_LIMIT_PER_HOUR:-}"
  # Per-user hourly ceiling for the document-AI actions (read/index, extract,
  # suggest, summarise — one shared bucket; optional; default 6, clamped
  # 1-1000). Raise on a self-host with a generous provider budget, lower on a
  # tight one. Listed here so an operator value in .env reaches the container
  # (the `environment:` block is a whitelist). See
  # docs/integrations/ai-providers.md.
  DOCUMENT_AI_LIMIT_PER_HOUR: "${DOCUMENT_AI_LIMIT_PER_HOUR:-}"
  # v1.17.1 — SMTP transport for the email notification channel (optional,
  # all-or-none: HOST + PORT + FROM enable it; USER/PASS optional; SECURE
  # toggles implicit TLS). Listed here so operator values in .env reach the
  # container (the `environment:` block is a whitelist — vars not listed
  # never propagate from .env, even with `${VAR}` substitution at
  # compose-up). Per-user recipient is stored in the DB, not here.
  SMTP_HOST: "${SMTP_HOST:-}"
  SMTP_PORT: "${SMTP_PORT:-}"
  SMTP_FROM: "${SMTP_FROM:-}"
  SMTP_USER: "${SMTP_USER:-}"
  SMTP_PASS: "${SMTP_PASS:-}"
  SMTP_SECURE: "${SMTP_SECURE:-}"
  # v1.17.1 — Web Push VAPID keys (optional). The admin Settings panel stores
  # these in the DB (the supported path); these env vars are the fallback the
  # loader reads when the DB pair is unset. Listed here so the documented env
  # path reaches the container under the bundled compose. See
  # docs/self-hosting/notifications.md.
  VAPID_PUBLIC_KEY: "${VAPID_PUBLIC_KEY:-}"
  VAPID_PRIVATE_KEY: "${VAPID_PRIVATE_KEY:-}"
  VAPID_SUBJECT: "${VAPID_SUBJECT:-}"
  # APNs for a self-hosted native iOS build (optional, env-only — no admin UI
  # for the .p8). APNS_KEY_B64 is the escape-free base64 form preferred over
  # APNS_KEY/APNS_KEY_FILE; the key must be scoped "Sandbox & Production".
  # PWA users need none of this (Web Push covers them). See
  # docs/self-hosting/notifications.md.
  APNS_KEY_B64: "${APNS_KEY_B64:-}"
  APNS_KEY_ID: "${APNS_KEY_ID:-}"
  APNS_TEAM_ID: "${APNS_TEAM_ID:-}"
  APNS_BUNDLE_ID: "${APNS_BUNDLE_ID:-}"
  # Set "true" only if your own iOS build carries Apple's Critical Alerts
  # entitlement; otherwise urgent alerts use the time-sensitive level.
  APNS_CRITICAL_ENTITLEMENT: "${APNS_CRITICAL_ENTITLEMENT:-}"

services:
  # Single-container default (HEALTHLOG_PROCESS_TYPE=all). Existing
  # self-host deployments stay working with `docker compose up`.
  # For horizontal scaling switch to the web/worker split below
  # (HEALTHLOG_PROCESS_TYPE=web on this service + uncomment app-worker).
  app:
    # Tracks the GHCR `:latest` multi-arch image by default.
    # `pull_policy: always` is the load-bearing line — without it,
    # Docker keeps a stale `:latest` digest cached on the host and
    # skips the registry round-trip on `compose up`, which is what kept
    # Coolify re-deploying the previous image after every tag push.
    # With pull_policy=always the compose-up always re-checks the
    # registry digest and pulls when GHCR has a fresher manifest.
    #
    # HEALTHLOG_IMAGE_REF pins the image deterministically without
    # editing this file: set it in the compose `.env` (or the Coolify
    # env vars) to `:vX.Y.Z` for a tag or `@sha256:<digest>` for an
    # immutable digest reference (see docs/ops/deploy.md, "Digest
    # pinning"; scripts/promote-digest.ts resolves + verifies the
    # digest after a release). Unset, everything behaves exactly as
    # before — `:latest` plus pull_policy.
    image: ghcr.io/mbombeck/healthlog${HEALTHLOG_IMAGE_REF:-:latest}
    pull_policy: always
    container_name: healthlog-app
    # The app runs as the unprivileged `nextjs` user and never needs a
    # setuid/setgid escalation; blocking privilege gain caps what a
    # compromised process inside the container can do. cap_drop removes
    # every Linux capability on top of that — the process doesn't need any
    # of them either. read_only:true is deliberately NOT set: the app
    # writes to a Next.js runtime cache, and turning the filesystem
    # read-only needs testing against that first, it's a not-yet, not an
    # oversight.
    security_opt:
      - no-new-privileges:true
    cap_drop:
      - ALL
    ports:
      - "3000:3000"
    environment:
      <<: *healthlog_env
    # Bring your own GeoLite2 databases. The published image only carries
    # MMDBs if the build had a MaxMind licence key, so on `:latest` the
    # offline tier is usually absent and every login IP goes to the online
    # provider instead. Create a free MaxMind account, pull
    # GeoLite2-City.mmdb + GeoLite2-ASN.mmdb with `geoipupdate`, mount the
    # directory read-only, and uncomment below — the container path here
    # matches GEOLITE2_DIR's default (/opt/geolite2), so this works as-is
    # once uncommented. Mounting at a different container path also works,
    # just also set GEOLITE2_DIR in .env to match it.
    # Steps: docs/self-hosting/geolite2.md.
    #
    # volumes:
    #   - ./geolite2:/opt/geolite2:ro
    depends_on:
      db:
        condition: service_healthy
    restart: unless-stopped
    # Bounded log rotation: the default json-file driver keeps every line
    # forever, which is exactly what fills the disk on the small NAS /
    # homelab hosts this file targets. 10m x 3 caps it at 30MB.
    logging:
      driver: json-file
      options:
        max-size: "10m"
        max-file: "3"
    # Conservative memory ceiling for a small host. HEALTHLOG_PROCESS_TYPE=all
    # bundles the AI/rollup worker into this same process, and unbounded that
    # workload can grow large enough under load to starve Postgres of memory
    # on a shared host. 1g is a starting point, not a measured ceiling — raise
    # it with APP_MEMORY_LIMIT in .env if the app gets OOM-killed under real
    # load.
    mem_limit: "${APP_MEMORY_LIMIT:-1g}"
    healthcheck:
      test:
        [
          "CMD-SHELL",
          "wget --no-verbose --tries=1 --spider http://127.0.0.1:3000/api/health || exit 1",
        ]
      interval: 30s
      timeout: 5s
      retries: 3
      start_period: 30s

  # Optional dedicated worker container — uncomment + set
  # HEALTHLOG_PROCESS_TYPE=web on `app` to split the deployment.
  # See docs/self-hosting/scaling.md for the full guide.
  #
  # app-worker:
  #   image: ghcr.io/mbombeck/healthlog${HEALTHLOG_IMAGE_REF:-:latest}
  #   build:
  #     context: .
  #     dockerfile: Dockerfile
  #   container_name: healthlog-worker
  #   security_opt:
  #     - no-new-privileges:true
  #   cap_drop:
  #     - ALL
  #   environment:
  #     <<: *healthlog_env
  #     HEALTHLOG_PROCESS_TYPE: "worker"
  #   depends_on:
  #     db:
  #       condition: service_healthy
  #   restart: unless-stopped
  #   logging:
  #     driver: json-file
  #     options:
  #       max-size: "10m"
  #       max-file: "3"
  #   mem_limit: "${WORKER_MEMORY_LIMIT:-1g}"

  db:
    # Pinned to the concrete minor instead of the floating `16-alpine`
    # alias so a database update is always a deliberate, reviewable
    # change (at pin time `16-alpine` resolved to this same digest, so
    # existing deployments see no image change). Minor releases share
    # the on-disk format — bumping this pin is safe without a dump /
    # restore; a major-version bump (17+) is not.
    image: postgres:16.14-alpine
    container_name: healthlog-db
    # cap_drop: [ALL] is deliberately NOT applied here (unlike `app`): the
    # official Postgres entrypoint starts as root, chowns the data
    # directory, and drops privileges to the `postgres` user itself, which
    # needs CAP_CHOWN/CAP_SETUID/CAP_SETGID to do. Dropping all capabilities
    # makes that startup fail outright (verified locally — the container
    # exits with "failed switching to 'postgres': operation not permitted").
    # no-new-privileges still blocks privilege escalation beyond that.
    security_opt:
      - no-new-privileges:true
    environment:
      POSTGRES_USER: healthlog
      POSTGRES_PASSWORD: "${POSTGRES_PASSWORD}"
      POSTGRES_DB: healthlog
    volumes:
      - pgdata:/var/lib/postgresql/data
    healthcheck:
      test: ["CMD-SHELL", "pg_isready -U healthlog"]
      interval: 5s
      timeout: 5s
      retries: 5
    restart: unless-stopped
    # No `mem_limit` here, deliberately, and it is not an oversight. Bounding
    # Postgres is the more dangerous half of the pair: an app container the
    # kernel kills restarts clean, a database killed mid-write does not. With
    # only `app` bounded the app is what gives way first under host pressure,
    # which is the failure order we want. Raise APP_MEMORY_LIMIT before
    # reaching for a limit here.
    # Bounded log rotation, same reasoning as `app`: default json-file grows
    # without bound and fills the disk on a small NAS/homelab host.
    logging:
      driver: json-file
      options:
        max-size: "10m"
        max-file: "3"

volumes:
  pgdata: