Environment variables

Claudie services are configured via environment variables. Most of them are wired through a ConfigMap named env, which is generated from the .env file shipped with the Claudie manifests.

How the variables reach the services:

• Each Deployment references specific keys of the env ConfigMap explicitly (there is no envFrom), so adding a new key to the ConfigMap alone has no effect unless the Deployment also references it.
• Any variable that is not set falls back to the default value built into the service (listed below).
• DATABASE_URL (manager) and AWS_ACCESS_KEY_ID/AWS_SECRET_ACCESS_KEY (terraformer) are sourced from the mongo-secret and minio-secret Secrets, not from the ConfigMap.
• NAMESPACE is injected automatically from the pod metadata, and NATS_CLUSTER_URL is set directly in each Deployment to nats://nats.$(NAMESPACE).svc.cluster.local.
• To set a variable that is not wired into a Deployment (e.g. CLAUDIE_NAMESPACES or PING_CONCURRENT_WORKERS), add it to the respective Deployment.

Changes to the ConfigMap are reflected after the respective services are restarted.

Common

Variables shared by all services. They are read by every service, though some only take effect in a specific service, noted in the description.

Variable	Default	Description
GOLANG_LOG	info	Log level, one of disabled, panic, fatal, error, warn, info, debug, trace.
NATS_CLUSTER_URL	nats://127.0.0.1	URL of the NATS cluster used for inter-service messaging. Set by each Deployment to nats://nats.$(NAMESPACE).svc.cluster.local.
NATS_CLUSTER_PORT	4222	Port appended to NATS_CLUSTER_URL.
NATS_CLUSTER_SIZE	1	Expected size of the NATS cluster when connecting. The shipped .env sets 3.
NATS_CLUSTER_JETSTREAM_NAME	claudie-internal	Name of the JetStream work queue through which tasks are exchanged.
MANAGER_HOSTNAME	localhost	Hostname under which the manager gRPC API is reachable. Used by the manager's clients (claudie-operator, testing-framework) to connect. The shipped .env sets manager.
MANAGER_PORT	50055	Port of the manager gRPC API. The manager listens on it; clients use it together with MANAGER_HOSTNAME to connect.
DATABASE_URL	mongodb://localhost:27017	MongoDB connection string for storing Claudie configs. Sourced from the mongo-secret Secret. Only acted upon by the manager.
BUCKET_URL	none	Endpoint of the MinIO/S3 deployment storing the tofu state files. When empty, an external AWS S3 bucket is used based on AWS_REGION and BUCKET_NAME. The shipped .env sets http://minio:9000. Only acted upon by terraformer.
BUCKET_NAME	claudie-tf-state-files	Name of the bucket holding the tofu state files. Only acted upon by terraformer.
AWS_REGION	local	Region of the bucket. Only acted upon by terraformer.
AWS_ACCESS_KEY_ID	fake	Access key for the bucket. Sourced from the minio-secret Secret. Only acted upon by terraformer.
AWS_SECRET_ACCESS_KEY	fake	Secret key for the bucket. Sourced from the minio-secret Secret. Only acted upon by terraformer.
PROMETHEUS_PORT	9090	Port of the /metrics HTTP endpoint.
NAMESPACE	none	Namespace of the Claudie deployment, injected automatically from the pod metadata. When empty, the services assume a local (non-Kubernetes) deployment.

Manager

Variable	Default	Description
MANAGER_TICK_FOR_INFRA_REFRESH	100	Number of idle reconciliation ticks before a full infrastructure refresh is scheduled (100 ticks ≈ 35 minutes).
MANAGER_TIME_FOR_NODE_DELETION	10	How many minutes a node may stay unreachable before Claudie deletes and replaces it.
PING_CONCURRENT_WORKERS	10	How many concurrent workers ssh ping the loadbalancer nodes during healthchecks.

Terraformer

Variable	Default	Description
TERRAFORMER_PORT	50052	Port of the terraformer gRPC healthcheck server.
TERRAFORMER_CONCURRENT_PROCESSES	8	How many tofu processes terraformer runs concurrently.
TERRAFORMER_TOFU_PARALLELISM	40	How many resources tofu works on in parallel (--parallelism).

Ansibler

Variable	Default	Description
ANSIBLER_PORT	50053	Port of the ansibler gRPC healthcheck server.
ANSIBLER_CONCURRENT_CLUSTERS	8	How many clusters ansibler works on concurrently.
ANSIBLER_FORKS	32	With how many forks each ansible playbook is spawned (--forks).

Kube-eleven

Variable	Default	Description
KUBE_ELEVEN_PORT	50054	Port of the kube-eleven gRPC healthcheck server.
KUBE_ELEVEN_CONCURRENT_CLUSTERS	7	How many clusters kube-eleven works on concurrently.

Kuber

Variable	Default	Description
KUBER_PORT	50057	Port of the kuber gRPC healthcheck server.
KUBER_CONCURRENT_WORKERS	30	How many nodes are patched concurrently.
KUBER_CONCURRENT_CALLS	90	How many kubectl processes kuber runs concurrently.

Claudie-operator

Variable	Default	Description
OPERATOR_PORT	50058	Port of the claudie-operator gRPC server.
CLAUDIE_NAMESPACES	all namespaces	Comma-separated list of namespaces in which InputManifests are watched for. See deploying Claudie in a custom namespace.
WEBHOOK_TLS_PORT	9443	TLS port of the InputManifest validating webhook.
WEBHOOK_CERT_DIR	./tls	Directory with the webhook server key and certificate. The shipped manifest sets /etc/webhook/certs/.
WEBHOOK_PATH	/validate-manifest	HTTP path served by the validating webhook.
CLAUDIE_TEMPLATES_REFERENCE_HTTPS_URL	github.com/berops/claudie-config	Repository of the default claudie-default-templates TemplateGitReference created by the operator.
CLAUDIE_TEMPLATES_REFERENCE_COMMIT	release	Ref/commit of the default TemplateGitReference.

Advanced service internals

Each of the manager, terraformer, ansibler, kube-eleven and kuber services additionally reads the following variables, where <SERVICE> is the service name (e.g. KUBE_ELEVEN_ACK_WAIT_TIME). These rarely need to be changed.

Variable	Default	Description
<SERVICE>_DURABLE_NAME	service name	Name of the durable NATS JetStream consumer.
<SERVICE>_ACK_WAIT_TIME	8	How many minutes an in-flight task message may stay unacknowledged before NATS redelivers it.
<SERVICE>_HEALTHCHECK_READINESS_SERVICE_NAME	<service>-readiness	gRPC health service name used by the readiness probe.
<SERVICE>_HEALTHCHECK_LIVENESS_SERVICE_NAME	<service>-liveness	gRPC health service name used by the liveness probe.