### Global Options (all components) ### DEBUG=false # Enable verbose (debug-level) logging for every component (default: false) ### NATS Options ### NATS_URL=nats://localhost:4222 # NATS server URL (default: nats://localhost:4222) NATS_USERNAME= # NATS username (optional) NATS_PASSWORD= # NATS password (optional) NATS_TLS_ENABLED=false # Enable TLS for NATS connection (default: false) ### HadesAPI Options ### API_PORT=8080 # Port for the API to listen on (default: 8080) AUTH_KEY="" # Key used for Basic Auth with the API. REQUIRED: the API refuses to start without it unless ALLOW_UNAUTHENTICATED=true, and compose.yml refuses to start without it. Generate one with `openssl rand -hex 32`. ### Dashboard (operator web UI, served by HadesAPI) ### # The dashboard is disabled unless USERNAME, PASSWORD_HASH and SESSION_SECRET are # all set; when disabled its /api routes return 503 and the SPA is not served. DASHBOARD_USERNAME="" # Dashboard login username (default: "" = dashboard disabled) DASHBOARD_PASSWORD_HASH="" # bcrypt hash of the dashboard password (e.g. `htpasswd -bnBC 12 "" pass | tr -d ':\n'`) DASHBOARD_SESSION_SECRET="" # HMAC key signing session cookies, >=32 chars (use a long random string) DASHBOARD_SESSION_TTL=12h # Session lifetime as a Go duration (default: 12h) DASHBOARD_JOB_RETENTION=1h # How long finished jobs stay in the dashboard list (default: 1h) DASHBOARD_TRUSTED_PROXIES= # Comma-separated CIDRs/IPs of front proxies allowed to set X-Forwarded-For (empty = trust none; login lockout keys on the direct address) DASHBOARD_COOKIE_INSECURE=false # Drop the Secure cookie flag - LOCAL HTTP DEV ONLY; requires TLS in real deployments LOG_MANAGER_URL=http://hades-log-manager-service:8081 # Internal HadesLogManager URL for the logs proxy SECRET_REDACT_MODE=smart # Metadata redaction: "smart" (key+value heuristics) or "all" (mask every value) SECRET_KEY_PATTERNS= # Optional regex overriding which metadata keys are masked (default covers token/password/secret/...) # The following four variables are reserved and NOT currently read by any component. # They are kept for forward compatibility; setting them has no effect today. PROMETHEUS_ADDRESS="" # Reserved: address for Prometheus metrics (not yet implemented) RETENTION_IN_MIN=30 # Reserved: minutes to keep completed jobs for monitoring (not yet implemented) MAX_RETRIES=3 # Reserved: retries before a job is marked failed (not yet implemented) TIMEOUT_IN_MIN=0 # Reserved: max task runtime before kill, 0 = no limit (not yet implemented) ### Tracing (all components) ### OTEL_EXPORTER_OTLP_ENDPOINT="" # OTLP gRPC endpoint for OpenTelemetry traces (e.g. http://jaeger:4317). Empty = tracing disabled (noop). # OTEL_SERVICE_NAME="" # Optional: override the service name shown in traces (defaults per service). ### HadesExecutor Options ### HADES_EXECUTOR="docker" # Executor to use for running jobs (options: docker, k8s; default: docker) ### HadesScheduler Options ### CONCURRENCY=1 # Number of jobs to run concurrently (default: 1) NATS_ACK_WAIT=1m # How long JetStream waits for an ack before redelivering a job to another worker; a liveness backstop, not a job-duration budget (default: 1m, minimum: 300ms) NATS_MAX_DELIVER=3 # Max deliveries per job; the last one publishes a terminal Failed status instead of running the job again (minimum: 2, default: 3) ### HadesOperator Options ### WATCH_NAMESPACE="hades" DELETE_ON_COMPLETE="true" MAX_PARALLELISM="100" REQUEUE_DELAY="2s" # How often the operator re-reconciles a running BuildJob (Go duration, default: 2s) LOG_DRAIN_TIMEOUT="45s" # How long a completed BuildJob is kept while its logs drain (Go duration, default: 45s) FINALIZER_IMAGE="busybox:1.37.0" # Image for the main container of every BuildJob pod (it only runs `echo build finished`). Keep it tagged: an untagged image means :latest, which Kubernetes forces to imagePullPolicy: Always (default: busybox:1.37.0) ### HadesLogManager Options ### HADESLOGMANAGER_API_PORT=8081 # Port for the Log Manager to listen on (default: 8081) LOG_BATCH_SIZE=100 # Log entries buffered before a flush (default: 100) LOG_RETENTION=1h # How long completed job logs are kept in memory (Go duration, default: 1h) MAX_JOB_LOGS=1000 # Max log entries retained per job (default: 1000) # Where aggregated logs are forwarded is now configured per job via the "callback_url" field in the build request, not globally. STATUS_WEBHOOK_ENABLED=true # Deliver the job-status webhook to each job's "status_callback_url" (default: true) STATUS_WEBHOOK_MAX_ATTEMPTS=6 # Delivery attempts per job before the status event is dropped (default: 6) STATUS_WEBHOOK_TIMEOUT=10s # Bound on a single status webhook delivery (default: 10s) STATUS_WEBHOOK_INITIAL_BACKOFF=5s # Delay before the second attempt; doubles per attempt (default: 5s) STATUS_WEBHOOK_MAX_BACKOFF=5m # Ceiling for the retry delay (default: 5m) STATUS_WEBHOOK_CONCURRENCY=16 # Status webhook deliveries in flight at once (default: 16) STATUS_WEBHOOK_MAX_PENDING=1000 # Status events awaiting acknowledgement, in flight or backing off (default: 1000) # The status webhook destination is per job via the "status_callback_url" field in the build request; there is no global URL. ### DockerScheduler Options ### DOCKER_HOST="unix:///var/run/docker.sock" # Docker host to connect to (default: unix:///var/run/docker.sock) DOCKER_SCRIPT_EXECUTOR="/bin/bash -c" # Path to the script executor to use (default: /bin/sh) DOCKER_CPU_LIMIT=2 # Number of CPUs to limit the container to (see Docker documentation for more info) DOCKER_MEMORY_LIMIT=2g # Memory limit for the container (see Docker documentation for more info) DOCKER_CONTAINER_AUTOREMOVE=false # Disable auto-removal to retain container logs post-execution (default: false) DOCKER_IMAGE_PULL_POLICY=always # always | if-not-present | never. "always" contacts the registry once per step even when the image is already local, which costs ~2.2s per step for a small image and more for a large one; "if-not-present" skips that but never picks up a moved tag (default: always) DOCKER_IMAGE_PULL_FRESHNESS=0s # under "always", skip the registry for a present image this scheduler pulled within this window, e.g. 30s (default: 0s) ### Traefik Deployment configuration ### LETSENCRYPT_EMAIL=you@example.com HADES_API_HOST=hades.example.com