Examples

Prerequisites

These walkthroughs assume a running Kollect operator and kubectl access. Start with QUICKSTART.md or Kind local lab if you have not installed the controller yet.

Walkthroughs backed by config/samples/. Apply defaults:

kubectl apply -k config/samples/

Example	Topic
Deployment inventory	Profile → Target → Inventory → Postgres 30s + Git 1h dual-cadence (ADR-0413)
Helm / Argo release inventory	Argo CD Application (Flux secondary)
Multi-cluster fleet	N operators → shared Postgres/Git (ADR-0501)
Postgres state store	Relational SoR + delete reconciliation
NATS event sink	JetStream events
Kafka event sink	Kafka topic export
Cluster-scoped rollup	Cluster CRDs + dedupe
Multi-tenant watch scope	Scope + watchNamespaces
Team-owned operator	values-minimal-rbac.yaml + config/samples/team-operator/
Connection test	KollectConnectionTest workflow
Cert-manager webhooks	Webhook TLS install
Kind local lab	kind quickstart
Wide-scope kind demo	Sales-pitch showcase — Trivy, certs, Git export, UI reveal
Wide-scope demo ROADMAP	Personas, churn presets, early-adopter checklist
UI local development	Mock vs live Read API for kollect-ui

Samples not in default kustomization

NATS, Kafka, and some cluster-scoped samples are documented but not included in kubectl apply -k config/samples/. Apply those files individually when following their guides.

S3/GCS Parquet — not sampled yet

JSON object export is shipped for type: s3 and type: gcs. DuckDB-queryable Parquet snapshot mode is planned per ADR-0401 — there is no kustomized sample under config/samples/ yet. Use the JSON sinks in postgres-state-store.md or deployment-inventory for end-to-end export until a Parquet walkthrough lands.

Per-sink export intervals

Dual-cadence Postgres + Git is wired in config/samples/kollect_v1alpha1_kollectinventory.yaml and explained in deployment-inventory.md (ADR-0413).