Skip to content

For the complete documentation index, see llms.txt.

Deployment configuration

This reference lists the PIGDeployment fields and pig-supervisor bootstrap chart values used by the default Helm installation. It covers resource identity, the collector Ingress, release policy, analyzer pod settings, the hosted connection, storage, and analysis. For a manually managed installation, see the manual Helm reference, which covers the worker chart 0.3.6 values. Those values are a separate interface.

apiVersion: governance.promptless.ai/v1alpha1
kind: PIGDeployment
metadata:
name: acme
namespace: pig

Each PIGDeployment identifies itself with these fields:

FieldMeaning
apiVersiongovernance.promptless.ai/v1alpha1.
kindPIGDeployment.
metadata.nameAt most 54 lowercase letters, digits, or hyphens, starting with a letter and ending with a letter or digit. The generated analyzer Deployment and Service add -analyzer to this name, such as acme-analyzer.
metadata.namespaceThe namespace the supervisor watches, set by the bootstrap chart’s watchNamespace. Secret and ConfigMap references resolve here.

Keep the resource name, namespace, and registered deployment ID stable. Create exactly one PIGDeployment in each watched namespace. Install one supervisor for that namespace, and keep its bootstrap values with your recovery records.

The pig-supervisor bootstrap chart creates the HTTPS collector Ingress. Configure it in the chart’s ingress values, separately from the PIGDeployment resource. The Ingress is created in watchNamespace and routes the collector paths to the analyzer Service.

ValueMeaning
watchNamespaceRequired. Namespace of the PIGDeployment, its Secrets, and the analyzer, such as pig. The Ingress is created there.
image.digestSupervisor image digest. The published chart sets it; carry forward the running digest before a later helm upgrade.
releaseCatalogURLRelease catalog the supervisor reads. Defaults to the public stable catalog.
ingress.serviceNameRequired. The analyzer Service name, your PIGDeployment name followed by -analyzer, such as acme-analyzer. Also names the Ingress.
ingress.hostnameRequired. Analyzer hostname reachable from enrolled hosts.
ingress.ingressClassNameRequired. An existing Kubernetes ingress class.
ingress.tlsSecretNameOptional TLS Secret covering the hostname; omit when the ingress controller supplies the certificate, such as ACM.
ingress.annotationsController-specific annotations, such as the ingress-nginx request-body limit or AWS ALB scheme and ACM certificate.

The Ingress belongs to the bootstrap Helm release. Deleting or pausing the PIGDeployment leaves it in place; uninstalling the bootstrap release removes it. Ingress changes do not change the analyzer’s registered configuration or trigger its revalidation. To change them, follow change the collector ingress.

Fields are under spec.release:

FieldMeaning
channelstable follows every stable release, including major releases and schema migrations.
pausedfalse permits release transitions. true stops new transitions while preserving service and credential refresh.
pinnedVersionEmpty follows the channel; an exact supported release pins the target. A pause takes precedence.
confirmation.configMapRefName of a customer-owned ConfigMap in the deployment namespace with confirmations for the exact deployment and target release.

The policy coordinates the analyzer and supervisor together. A pin does not bypass compatibility or migration checks. See updates and recovery.

The confirmation ConfigMap binds releaseDigest and the Kubernetes object’s deploymentUID. Recovery confirmation uses confirmedAt, postgresRecoveryPoint, and objectRecoveryPoint. Operator capacity acknowledgement uses capacityConfirmedAt, capacityRequirementsDigest, and capacityEvidence. Supply the fields required by the target release; see confirm release prerequisites. These acknowledgements do not replace live dependency checks or mirror Terraform state.

Fields are under spec:

FieldMeaning
serviceAccountNameExisting analyzer ServiceAccount, such as pig-analyzer, with cloud data-access bindings.
podLabelsAdditional analyzer pod labels required by your platform, such as Azure’s workload identity opt-in.
nodeSelectorOptional node labels that the analyzer pod and its preflight and migration Jobs must match. On GKE Standard, set iam.gke.io/gke-metadata-server-enabled: "true". Omit it on GKE Autopilot, which rejects it; see Deploy on GCP.
hosted.runtimeURLOptional Promptless endpoint override. Defaults to https://api.gopromptless.ai.
hosted.installTokenSecretRefSecret name and key containing the deployment token.

The generated Service uses port 8080. The supervisor derives the registered configuration from the effective settings. The install token identifies the installation and is separate from individual host credentials.

Fields are under spec.storage. Configure postgres and exactly one of s3, azureBlob, or gcs.

FieldMeaning
postgres.dsnSecretRefSecret name and key containing the database connection string with TLS settings.
postgres.migrationDsnSecretRefOptional Secret name and key for a separate migration credential. The supervisor passes it to preflight and migration Jobs; the serving analyzer uses postgres.dsnSecretRef.
postgres.caConfigMapRefOptional ConfigMap name and key for a CA bundle mounted at /etc/pig/postgres-ca/ca.pem.
s3.region, s3.bucket, s3.prefixAWS region, bucket, and dedicated trace prefix. Uses the analyzer’s AWS workload identity.
azureBlob.accountURL, azureBlob.container, azureBlob.prefixHTTPS storage account URL, private container, and trace prefix. Uses Microsoft Entra Workload ID.
gcs.bucket, gcs.prefixGoogle Cloud Storage bucket and trace prefix. Uses Workload Identity Federation for GKE.

Storage configuration grants no cloud management authority. Changing a location does not move existing data. Plan that transition using the object-storage guidance.

Fields are under spec.analysis:

FieldMeaning
activationAtTimezone-aware ISO timestamp for analysis eligibility. Leave empty for ingestion only.
quietWindowHoursPositive interval after session activity before analysis; the example uses 0.5 hours.
model.provideropenai, azure_openai, or aws_bedrock.
model.authenticationapi_key, or aws_sigv4 for Bedrock.
model.baseURLA supported HTTPS Responses API endpoint.
model.nameThe model or deployment name available in your provider account.
model.apiKeySecretRefSecret name and key; omit for Signature Version 4.

The model settings are required even when activationAt is empty. An organization administrator selects the instruction repositories the analyzer reads in PIG Settings. The analyzer reads GitHub repositories with a main branch. Analysis and remediation work with GitHub repositories only, even when your hub publishes to another Git host.

For accepted endpoint shapes and authentication, see Model providers. Create Repo Issues and Propose Remediations use write access granted to the connected GitHub App for the selected repositories. A successful analysis clone does not verify that access.

Review configuration through its owner, such as your GitOps repository. Validate the manifest against the installed release’s custom resource definition:

Terminal window
kubectl apply --dry-run=server -f pig-deployment.yaml

Apply the change and inspect kubectl describe pigdeployment acme --namespace pig. Repeat deployment verification after changes to storage, identities, model access, or endpoints. Schema validation alone cannot prove those dependencies are reachable.