Files
ENI-projet-piscine/docker-compose.yml
T
Dorian 901ceffd72
Airflow / Lint et intégrité des DAGs (push) Successful in 1m10s
Airflow / Construction de l'image (push) Successful in 1m47s
fix(etl): fiabilise airflow-init, borne les DAGs ML et ajoute la CI Airflow
2026-09-21 11:18:02 +02:00

164 lines
6.4 KiB
YAML

# Piege : PGDATA de l'image timescaledb-ha vaut /home/postgres/pgdata/data, pas le chemin
# habituel de l'image postgres. Monte ailleurs, le volume ne retient rien, sans erreur.
# Piege : db/init est monte fichier par fichier. Monter le dossier masquerait les scripts
# d'init de l'image, dont timescaledb-tune. Ajouter un fichier impose une ligne ici.
name: enervision
# Piege : LocalExecutor fait tourner les taches comme sous-processus du scheduler, jamais du
# webserver. `airflow_ml_state` (modele entraine, magasin MLflow) n'a donc besoin d'etre monte
# que sur `airflow-scheduler` en pratique, mais reste partage avec le webserver pour que ce
# dernier puisse au besoin l'inspecter sans en devenir dependant.
x-airflow-common: &airflow-common
build:
context: .
dockerfile: etl/airflow/Dockerfile
environment: &airflow-common-env
AIRFLOW__CORE__EXECUTOR: LocalExecutor
AIRFLOW__CORE__LOAD_EXAMPLES: "false"
# Piege : pas de `:?` sur les secrets Airflow. Compose interpole le fichier entier avant de
# filtrer les services : une variable requise manquante casserait aussi `make db-up`,
# `make dev`... pour quiconque n'a pas encore complete son `.env`. Le refus est porte par
# `airflow-init` (ci-dessous), dont `webserver` et `scheduler` dependent.
AIRFLOW__CORE__FERNET_KEY: ${AIRFLOW_FERNET_KEY:-}
AIRFLOW__WEBSERVER__SECRET_KEY: ${AIRFLOW_WEBSERVER_SECRET_KEY:-}
AIRFLOW__DATABASE__SQL_ALCHEMY_CONN: postgresql+psycopg2://${POSTGRES_USER}:${POSTGRES_PASSWORD}@db:5432/airflow
# Role `enervision_ml` dedie pas encore provisionne (dette assumee, cf. ADR 0003) :
# memes identifiants que le backend en attendant.
ML_DATABASE_URL: postgresql+psycopg://${POSTGRES_USER}:${POSTGRES_PASSWORD}@db:5432/${POSTGRES_DB}
MLFLOW_TRACKING_URI: sqlite:////opt/ml/state/mlflow.db
volumes:
- ./etl/airflow/dags:/opt/airflow/dags
- ./etl/airflow/plugins:/opt/airflow/plugins
- airflow_logs:/opt/airflow/logs
- airflow_ml_state:/opt/ml/state
restart: unless-stopped
services:
db:
image: timescale/timescaledb-ha:pg17
environment:
POSTGRES_USER: ${POSTGRES_USER:?}
POSTGRES_PASSWORD: ${POSTGRES_PASSWORD:?}
POSTGRES_DB: ${POSTGRES_DB:?}
TIMESCALEDB_TELEMETRY: ${TIMESCALEDB_TELEMETRY:-off}
ports:
- "${POSTGRES_PORT:-5433}:5432"
volumes:
- pgdata:/home/postgres/pgdata/data
- ./db/init/100-extensions.sql:/docker-entrypoint-initdb.d/100-extensions.sql:ro
- ./db/init/110-test-database.sql:/docker-entrypoint-initdb.d/110-test-database.sql:ro
- ./db/init/120-airflow-database.sql:/docker-entrypoint-initdb.d/120-airflow-database.sql:ro
healthcheck:
test: ["CMD-SHELL", "pg_isready -U $${POSTGRES_USER} -d $${POSTGRES_DB}"]
interval: 10s
timeout: 5s
retries: 12
start_period: 40s
restart: unless-stopped
# Piege : Mailpit ne relaie rien vers l'exterieur, il capture tout email envoye par le
# backend. Aucun acces reseau sortant n'est requis ; l'UI web (8025) sert a lire les emails.
mailpit:
image: axllent/mailpit
ports:
- "${MAILPIT_SMTP_PORT:-1025}:1025"
- "${MAILPIT_UI_PORT:-8025}:8025"
restart: unless-stopped
backend:
build: ./apps/backend
depends_on:
db:
condition: service_healthy
mailpit:
condition: service_started
environment:
APP_ENV: ${APP_ENV:-local}
APP_DEBUG: ${APP_DEBUG:-false}
APP_LOG_LEVEL: ${APP_LOG_LEVEL:-INFO}
APP_SECRET_KEY: ${APP_SECRET_KEY:?}
APP_CORS_ORIGINS: ${APP_CORS_ORIGINS:-http://localhost:4200}
DATABASE_URL: postgresql+asyncpg://${POSTGRES_USER}:${POSTGRES_PASSWORD}@db:5432/${POSTGRES_DB}
APP_MOCK_API_BASE_URL: ${APP_MOCK_API_BASE_URL:-https://api-mock.charlieandre.fr}
APP_MOCK_API_USERNAME: ${APP_MOCK_API_USERNAME:-}
APP_MOCK_API_PASSWORD: ${APP_MOCK_API_PASSWORD:-}
APP_MOCK_API_TIMEOUT_SECONDS: ${APP_MOCK_API_TIMEOUT_SECONDS:-10}
APP_FRONTEND_RESET_PASSWORD_URL: ${APP_FRONTEND_RESET_PASSWORD_URL:-http://localhost:4200/reset-password}
APP_SMTP_HOST: mailpit
APP_SMTP_PORT: "1025"
APP_SMTP_USE_TLS: "false"
APP_SMTP_FROM_ADDRESS: ${APP_SMTP_FROM_ADDRESS:-no-reply@enervision.fr}
ports:
- "${BACKEND_PORT:-8000}:8000"
restart: unless-stopped
frontend:
build: ./apps/frontend
ports:
- "${FRONTEND_PORT:-3000}:80"
restart: unless-stopped
# Conteneur unique, jamais redemarre. La migration et la creation du premier compte sont
# portees par l'entrypoint de l'image (`_AIRFLOW_DB_MIGRATE`, `_AIRFLOW_WWW_USER_*`), qui porte
# aussi leur code de sortie : une migration ratee (ex. base `airflow` absente sur un volume
# `pgdata` deja peuple) fait echouer ce service, et `webserver`/`scheduler`, qui attendent son
# succes, ne demarrent pas sur une base non migree. Le mot de passe passe par l'environnement,
# jamais par `argv` (ni `ps`, ni `docker compose config`).
# Sans mot de passe, l'entrypoint refuse lui-meme de creer le compte ; la commande ci-dessous
# refuse en plus les deux cles de chiffrement vides.
airflow-init:
<<: *airflow-common
restart: "no"
environment:
<<: *airflow-common-env
_AIRFLOW_DB_MIGRATE: "true"
_AIRFLOW_WWW_USER_CREATE: "true"
_AIRFLOW_WWW_USER_USERNAME: ${AIRFLOW_ADMIN_USERNAME:-admin}
_AIRFLOW_WWW_USER_PASSWORD: ${AIRFLOW_ADMIN_PASSWORD:-}
_AIRFLOW_WWW_USER_EMAIL: ${AIRFLOW_ADMIN_EMAIL:-admin@enervision.fr}
depends_on:
db:
condition: service_healthy
command:
- bash
- -c
- |
set -euo pipefail
: "$${AIRFLOW__CORE__FERNET_KEY:?AIRFLOW_FERNET_KEY manquant dans .env}"
: "$${AIRFLOW__WEBSERVER__SECRET_KEY:?AIRFLOW_WEBSERVER_SECRET_KEY manquant dans .env}"
exec airflow version
airflow-webserver:
<<: *airflow-common
command: webserver
ports:
- "${AIRFLOW_PORT:-8080}:8080"
depends_on:
db:
condition: service_healthy
airflow-init:
condition: service_completed_successfully
healthcheck:
test: ["CMD", "curl", "--fail", "http://localhost:8080/health"]
interval: 30s
timeout: 10s
retries: 5
start_period: 60s
airflow-scheduler:
<<: *airflow-common
command: scheduler
depends_on:
db:
condition: service_healthy
airflow-init:
condition: service_completed_successfully
volumes:
pgdata:
airflow_logs:
airflow_ml_state: