Platform catalog¶
73 curated open-source components in 10 groups, installed with Helm or kustomize by cloudseed platform install <group|item> with values adapted to the target and distro. Versions are pinned per cloudseed release; --version overrides one item. Dependencies are installed first and never twice.
Generated from the code
This page is generated by scripts/gen-docs.py from the catalog in cloudseed/platform.py, so it always matches the CLI you install (this build: cloudseed 0.2.3).
Legend: core items are installed by the group, extra items only by name. FIPS: compatible items install in a FIPS environment, tls-restricted and crypto-restricted items install and are flagged by cloudseed scan fips, items marked no are refused there unless --force (FIPS 140 mode).
Groups¶
| Group | What it brings | Core | Extras |
|---|---|---|---|
basek8s | Cluster management: metrics, certs, Gateway API + Envoy Gateway (private LB), GitOps, observability (Prometheus, Loki), OTel, secrets | 14 | 5 |
scaling | Autoscaling: KEDA, VPA + Goldilocks (+ metrics-server), cluster-autoscaler / Karpenter (cloud) | 4 | 1 |
data | Data platform: MinIO, Postgres operator, Kafka, Spark, Trino, StarRocks, Polaris, Airflow, ClickHouse | 9 | 1 |
ai | AI/ML: Ray, Kubeflow trainer/pipelines, KServe, JupyterHub, MLflow, vLLM, GPU operator, Ollama | 5 | 4 |
agentic | Agentic: kagent, kmcp, agentgateway, Qdrant, Langfuse, LiteLLM, Open WebUI | 4 | 3 |
finops | FinOps: OpenCost (+ Prometheus), kube-green, Kubecost; pairs with cs finops for cloud bills and estimates | 1 | 2 |
devsecops | DevSecOps: GitLab + runner (once GITLAB_RUNNER_TOKEN is set) + CI template, NeuVector, Trivy operator, ArgoCD; registries: Harbor, Artifactory, Nexus; SonarQube | 4 | 4 |
security | Security / zero trust: Istio (ambient or sidecar, strict mTLS), Falco, Kyverno policies, cert-manager + cluster CA issuer, External Secrets; extras: gateway, Kiali, Vault, Kubescape operator | 3 | 4 |
resilience | Resilience / DR: Velero backups (cloud bucket + identity created on demand), kured reboots, descheduler; cs dr runs backups, restores and drills | 3 | 0 |
chaos | Chaos engineering: Chaos Mesh (+ LitmusChaos extra); cs chaos run executes automated experiments with a PASS/FAIL report | 1 | 1 |
cloudseed platform list # the catalog with what is installed on the current cluster
cloudseed platform info basek8s # every tool a group brings
cloudseed platform plan basek8s # what an install would do
cloudseed platform install basek8s
basek8s¶
Cluster management: metrics, certs, Gateway API + Envoy Gateway (private LB), GitOps, observability (Prometheus, Loki), OTel, secrets
Install: cloudseed platform install basek8s · Explain: cloudseed explain group basek8s
| Item | What it is | Tier | Source | Targets | FIPS | Needs |
|---|---|---|---|---|---|---|
metrics-server | Resource metrics API (kubectl top, HPA) | core | metrics-server@3.14.0 | all | compatible | – |
cert-manager | X.509 certificates for in-cluster services (Gateway API aware) | core | cert-manager@v1.21.2 | all | crypto-restricted | gateway-api |
gateway-api | Kubernetes Gateway API CRDs (experimental channel v1.6.1, matches envoy-gateway) | core | https://github.com/kubernetes-sigs/gateway-api/releases/download/v1.6.1/experimental-install.yaml | all | compatible | – |
envoy-gateway | Envoy Gateway: the cluster's Gateway API implementation (private LB, TLS via cert-manager) | core | oci://docker.io/envoyproxy/gateway-helm@v1.9.1 | all | tls-restricted | gateway-api, cert-manager, cert-manager-issuer, metallb |
metallb | LoadBalancer IPs on the private network (local clusters) | core | metallb@0.16.1 | vmware | compatible | – |
argocd | GitOps continuous delivery | core | argo-cd@10.9.2 | all | compatible | – |
kube-prometheus-stack | Prometheus, Alertmanager, Grafana, node/kube exporters | core | kube-prometheus-stack@91.5.1 | all | compatible | – |
loki | Log aggregation (Loki 3, single binary, persistent) + Alloy log collection | core | loki@7.3.0 | all | compatible | alloy, local-path-provisioner (vmware) |
alloy | Grafana Alloy: collects every pod's logs into Loki | core | alloy@1.12.1 | all | compatible | – |
opentelemetry-operator | OpenTelemetry operator (collectors, auto-instrumentation) | core | opentelemetry-operator@0.123.1 | all | compatible | cert-manager |
external-secrets | Sync secrets from AWS/GCP/Azure secret managers, Vault, ... | core | external-secrets@2.11.0 | all | compatible | – |
sealed-secrets | Encrypt secrets so they can live in git | core | oci://registry-1.docker.io/bitnamicharts/sealed-secrets@2.20.0 | all | crypto-restricted | – |
cert-manager-issuer | Cluster CA issuer (cert-manager) for TLS on every UI | core | rendered by cloudseed | all | compatible | cert-manager |
local-path-provisioner | Default StorageClass for local clusters (Rancher local-path, node disks) | core | https://raw.githubusercontent.com/rancher/local-path-provisioner/v0.0.37/deploy/local-path-storage.yaml | vmware | compatible | – |
ingress-nginx | LEGACY Ingress controller (upstream retired 2026; use Gateway API / envoy-gateway) | extra | ingress-nginx@4.15.1 | all | tls-restricted | metallb |
aws-load-balancer-controller | ALB/NLB provisioning for EKS (needs IRSA role) | extra | aws-load-balancer-controller@3.5.0 | aws | compatible | – |
reloader | Restart workloads when ConfigMaps/Secrets change | extra | reloader@2.2.17 | all | compatible | – |
kyverno | Policy engine (admission control, mutation, generation) | extra | kyverno@3.9.1 | all | compatible | – |
external-dns | Publish Services/Ingresses/HTTPRoutes in Route53 / Cloud DNS / Azure DNS (cloud identity wired in) | extra | external-dns@1.22.0 | aws, gcp, azure | compatible | gateway-api |
scaling¶
Autoscaling: KEDA, VPA + Goldilocks (+ metrics-server), cluster-autoscaler / Karpenter (cloud)
Install: cloudseed platform install scaling · Explain: cloudseed explain group scaling
| Item | What it is | Tier | Source | Targets | FIPS | Needs |
|---|---|---|---|---|---|---|
keda | Event-driven autoscaling (queues, metrics, cron, ...) | core | keda@2.21.0 | all | compatible | – |
vpa | Vertical Pod Autoscaler | core | vpa@5.1.0 | all | compatible | metrics-server |
goldilocks | Right-sizing recommendations dashboard (uses VPA) | core | goldilocks@11.1.1 | all | compatible | vpa |
cluster-autoscaler | Scale node pools on demand within kubernetes_node_min/_max (EKS chart; GKE/AKS: built-in autoscaler, GKE never below kubernetes_node_count) | core | cluster-autoscaler@9.59.0 | aws, gcp, azure | compatible | – |
karpenter | Just-in-time nodes for EKS (controller + a default NodePool/EC2NodeClass) | extra | oci://public.ecr.aws/karpenter/karpenter@1.14.1 | aws | compatible | – |
data¶
Data platform: MinIO, Postgres operator, Kafka, Spark, Trino, StarRocks, Polaris, Airflow, ClickHouse
Install: cloudseed platform install data · Explain: cloudseed explain group data
| Item | What it is | Tier | Source | Targets | FIPS | Needs |
|---|---|---|---|---|---|---|
minio | S3-compatible object storage (lakehouse storage for local clusters) | core | minio@5.4.0 | all | compatible | local-path-provisioner (vmware) |
cloudnative-pg | PostgreSQL operator (CloudNativePG): Postgres clusters as CRDs (backs the Polaris catalog) | core | cloudnative-pg@0.29.1 | all | crypto-restricted | local-path-provisioner (vmware) |
strimzi | Apache Kafka operator (Strimzi) + a dev Kafka cluster (KRaft, 1 broker) | core | strimzi-kafka-operator@1.2.0 | all | no | local-path-provisioner (vmware) |
spark-operator | Kubeflow Spark operator (SparkApplication CRD) | core | spark-operator@2.5.2 | all | no | – |
trino | Distributed SQL query engine | core | trino@1.42.2 | all | no | – |
starrocks | StarRocks analytical database (operator + cluster CRD) | core | kube-starrocks@1.11.7 | all | no | local-path-provisioner (vmware) |
polaris | Apache Polaris - Iceberg REST catalog (Postgres-backed via CloudNativePG) | core | polaris@1.7.0 | all | no | cloudnative-pg, local-path-provisioner (vmware) |
spark-history-server | Spark History Server (event logs in MinIO s3a://spark-logs) | core | rendered by cloudseed | all | no | minio, spark-operator |
airflow | Apache Airflow (official chart) | core | airflow@1.22.0 | all | no | local-path-provisioner (vmware) |
clickhouse-operator | Altinity ClickHouse operator | extra | altinity-clickhouse-operator@0.27.3 | all | no | local-path-provisioner (vmware) |
ai¶
AI/ML: Ray, Kubeflow trainer/pipelines, KServe, JupyterHub, MLflow, vLLM, GPU operator, Ollama
Install: cloudseed platform install ai · Explain: cloudseed explain group ai
| Item | What it is | Tier | Source | Targets | FIPS | Needs |
|---|---|---|---|---|---|---|
kuberay | Ray operator (distributed Python, RayJob/RayService/RayCluster) | core | kuberay-operator@1.7.1 | all | no | – |
kubeflow-trainer | Kubeflow training operator (PyTorchJob, TFJob, ...) | core | github.com/kubeflow/training-operator/manifests/overlays/standalone?ref=v1.9.0 | all | no | – |
kserve | Model serving (KServe, standard (raw) deployment mode - no Knative needed) | core | oci://ghcr.io/kserve/charts/kserve-resources@v0.20.0 | all | no | cert-manager, kserve-crd |
jupyterhub | Multi-user notebooks | core | jupyterhub@4.4.2 | all | no | local-path-provisioner (vmware) |
mlflow | Experiment tracking and model registry | core | mlflow@1.11.7 | all | no | local-path-provisioner (vmware) |
kubeflow-pipelines | Kubeflow Pipelines (standalone, platform-agnostic) | extra | github.com/kubeflow/pipelines/manifests/kustomize/cluster-scoped-resources?ref=2.17.2 | all (amd64 only) | no | local-path-provisioner (vmware) |
vllm-stack | vLLM production stack (router + model servers) | extra | vllm-stack@0.1.12 | all (amd64 only) | no | – |
gpu-operator | NVIDIA GPU operator (drivers, device plugin, DCGM) | extra | gpu-operator@v26.7.1 | aws, gcp, azure | compatible | – |
ollama | Local LLM runtime (CPU-friendly for dev) | extra | ollama@1.83.0 | all | no | local-path-provisioner (vmware) |
agentic¶
Agentic: kagent, kmcp, agentgateway, Qdrant, Langfuse, LiteLLM, Open WebUI
Install: cloudseed platform install agentic · Explain: cloudseed explain group agentic
| Item | What it is | Tier | Source | Targets | FIPS | Needs |
|---|---|---|---|---|---|---|
agentgateway | agentgateway: gateway for agents, MCP and LLM traffic (GatewayClass agentgateway) | core | oci://cr.agentgateway.dev/charts/agentgateway@1.5.0 | all | no | gateway-api, agentgateway-crds |
kagent | kagent: agentic AI framework for Kubernetes (agents as CRDs) | core | oci://ghcr.io/kagent-dev/kagent/helm/kagent@0.10.1 | all | no | kagent-crds |
kmcp | kmcp: build and run MCP servers on Kubernetes | core | oci://ghcr.io/kagent-dev/kmcp/helm/kmcp@0.4.0 | all | no | kagent-crds |
qdrant | Vector database | core | qdrant@1.19.1 | all | no | local-path-provisioner (vmware) |
langfuse | LLM observability, tracing and evals | extra | langfuse@2.1.2 | all | no | clickhouse-operator-official, local-path-provisioner (vmware) |
litellm | LLM proxy / model router (OpenAI-compatible) | extra | oci://ghcr.io/berriai/litellm-helm@1.102.1 | all | no | – |
open-webui | Chat UI for local/remote models | extra | open-webui@16.6.0 | all | no | local-path-provisioner (vmware) |
finops¶
FinOps: OpenCost (+ Prometheus), kube-green, Kubecost; pairs with cs finops for cloud bills and estimates
Install: cloudseed platform install finops · Explain: cloudseed explain group finops
| Item | What it is | Tier | Source | Targets | FIPS | Needs |
|---|---|---|---|---|---|---|
opencost | OpenCost: real-time Kubernetes cost allocation (namespace, workload, label) | core | opencost@2.5.32 | all | compatible | kube-prometheus-stack |
kube-green | Sleep dev workloads on a schedule (kube-green) | extra | https://github.com/kube-green/kube-green/releases/download/v0.7.1/kube-green.yaml | all | compatible | cert-manager |
kubecost-cost-analyzer | Kubecost (free tier; richer UI on top of OpenCost) | extra | cost-analyzer@2.8.7 | all | no | local-path-provisioner (vmware) |
Also installed with this group (members of other groups): metrics-server, goldilocks.
devsecops¶
DevSecOps: GitLab + runner (once GITLAB_RUNNER_TOKEN is set) + CI template, NeuVector, Trivy operator, ArgoCD; registries: Harbor, Artifactory, Nexus; SonarQube
Install: cloudseed platform install devsecops · Explain: cloudseed explain group devsecops
| Item | What it is | Tier | Source | Targets | FIPS | Needs |
|---|---|---|---|---|---|---|
gitlab | GitLab (git, CI, registry) - dev-sized, self-signed ingress | core | gitlab@9.11.13 | all (amd64 only) | no | ingress-nginx, cert-manager, local-path-provisioner (vmware) |
gitlab-runner | GitLab CI runner (Kubernetes executor) | core | gitlab-runner@0.88.4 | all | no | gitlab |
neuvector | NeuVector container security platform (runtime, scanning, WAF) | core | core@2.11.2 | all | no | – |
trivy-operator | Continuous vulnerability/misconfiguration scanning of workloads | core | trivy-operator@0.36.0 | all | compatible | – |
artifactory | JFrog Artifactory OSS - universal artifact repository (Maven, npm, PyPI, generic) | extra | artifactory-oss@107.161.26 | all | no | local-path-provisioner (vmware) |
nexus | Sonatype Nexus Repository 3 - artifact + container registry | extra | nexus3@5.26.0 | all | no | local-path-provisioner (vmware) |
harbor | Container registry with scanning and signing | extra | harbor@1.19.2 | all (amd64 only) | no | local-path-provisioner (vmware) |
sonarqube | Code quality and SAST | extra | sonarqube@2026.4.1 | all | no | local-path-provisioner (vmware) |
Also installed with this group (members of other groups): argocd, kyverno.
security¶
Security / zero trust: Istio (ambient or sidecar, strict mTLS), Falco, Kyverno policies, cert-manager + cluster CA issuer, External Secrets; extras: gateway, Kiali, Vault, Kubescape operator
Install: cloudseed platform install security · Explain: cloudseed explain group security
| Item | What it is | Tier | Source | Targets | FIPS | Needs |
|---|---|---|---|---|---|---|
istio | Istio service mesh (mode: ambient [default] or sidecar) with strict mTLS | core | bundle: ambient = istio-base, istiod, istio-cni, ztunnel; sidecar = istio-base, istiod | all | no | – |
falco | Runtime threat detection (syscalls) + Falcosidekick UI | core | falco@9.2.0 | all | compatible | – |
kyverno-policies | Kyverno pod-security baseline policies (audit mode) | core | kyverno-policies@3.9.1 | all | compatible | kyverno |
istio-gateway | Istio ingress gateway (private LB address: VPN/bastion on clouds, host-only network on vmware) | extra | gateway@1.30.5 | all | no | istio |
kiali | Mesh observability console | extra | kiali-server@2.32.0 | all | no | istio |
vault | HashiCorp Vault (dev mode - not for production secrets) | extra | vault@0.34.1 | all | no | – |
kubescape-operator | Continuous configuration, vulnerability and node scanning (NSA / MITRE / CIS reports as CRDs) | extra | kubescape-operator@1.40.4 | all | compatible | – |
Also installed with this group (members of other groups): kyverno, cert-manager, cert-manager-issuer, external-secrets.
resilience¶
Resilience / DR: Velero backups (cloud bucket + identity created on demand), kured reboots, descheduler; cs dr runs backups, restores and drills
Install: cloudseed platform install resilience · Explain: cloudseed explain group resilience
| Item | What it is | Tier | Source | Targets | FIPS | Needs |
|---|---|---|---|---|---|---|
velero | Backup and restore of cluster state and volumes (Velero + node agent, cloud bucket or MinIO) | core | velero@12.2.0 | all | crypto-restricted | local-path-provisioner (vmware), minio (vmware) |
kured | Safe rolling node reboots when the OS asks for one (kured) | core | kured@6.1.0 | all | compatible | – |
descheduler | Rebalance pods across nodes (duplicates, node utilisation, affinity violations) | core | descheduler@0.36.0 | all | compatible | – |
chaos¶
Chaos engineering: Chaos Mesh (+ LitmusChaos extra); cs chaos run executes automated experiments with a PASS/FAIL report
Install: cloudseed platform install chaos · Explain: cloudseed explain group chaos
| Item | What it is | Tier | Source | Targets | FIPS | Needs |
|---|---|---|---|---|---|---|
chaos-mesh | Chaos Mesh: pod, network, DNS, stress, time and IO faults with a dashboard (cs chaos run drives it) | core | chaos-mesh@2.8.4 | all | compatible | – |
litmus | LitmusChaos ChaosCenter (portal + experiments hub) | extra | litmus@3.30.0 | all (amd64 only) | no | local-path-provisioner (vmware) |
Web UIs¶
cloudseed platform ui attaches an HTTPRoute on the shared private Gateway for every installed item below and prints its URL (https://<item>.<domain>) and login. Generated passwords live in <workdir>/platform/secrets.json (0600).
| Item | Namespace / service | Login |
|---|---|---|
airflow | airflow/airflow-api-server:8080 | admin / airflow_password (platform/secrets.json) |
argocd | argocd/argocd-server:80 | admin / kubectl -n argocd get secret argocd-initial-admin-secret -o jsonpath='{.data.password}' | base64 -d |
artifactory | artifactory/artifactory:8082 | admin / artifactory_password (platform/secrets.json) |
chaos-mesh | chaos-mesh/chaos-dashboard:2333 | token: cs kubectl create token cloudseed-chaos-admin -n chaos-mesh --duration=24h |
falco | falco/falco-falcosidekick-ui:2802 | admin / falco_ui_password (platform/secrets.json) |
gitlab | gitlab/gitlab-webservice-default:8181 | root / kubectl -n gitlab get secret gitlab-gitlab-initial-root-password -o jsonpath='{.data.password}' | base64 -d |
harbor | harbor/harbor:80 | admin / harbor_password (platform/secrets.json) |
jupyterhub | jupyterhub/proxy-public:80 | any user / jupyter_password (platform/secrets.json) |
kagent | kagent/kagent-ui:8080 | admin / kagent_password (platform/secrets.json) |
kiali | istio-system/kiali:20001 | anonymous |
kube-prometheus-stack | monitoring/monitoring-grafana:80 | admin / grafana_password (platform/secrets.json) |
kubecost-cost-analyzer | kubecost/kubecost-cost-analyzer:9090 | – |
kubeflow-pipelines | kubeflow/ml-pipeline-ui:80 | – |
langfuse | langfuse/langfuse-web:3000 | sign up on first visit |
litmus | litmus/litmus-frontend-service:9091 | admin / litmus (change it) |
minio | minio/minio-console:9001 | minio_root_user / minio_password (platform/secrets.json; admin when minio_root_user is not there: an install from before cloudseed generated it) |
mlflow | mlflow/mlflow:80 | – |
neuvector | neuvector/neuvector-service-webui:8443 | admin / admin (change it) |
nexus | nexus/nexus-nexus3:8081 | admin / nexus_password (platform/secrets.json) |
open-webui | open-webui/open-webui:80 | sign up on first visit |
opencost | opencost/opencost:9090 | – |
sonarqube | sonarqube/sonarqube-sonarqube:9000 | admin / admin (change it) |
spark-history-server | spark-operator/spark-history-server:18080 | – |
trino | trino/trino:8080 | any user, no password |
Step-overs and conflicts¶
cluster-autoscaleris built into gke, aks: cloudseed skips it there.metrics-serveris built into rke2, gke, aks: cloudseed skips it there.agentgatewayoverlapsenvoy-gateway: both can run, the plan warns.artifactoryoverlapsharbor,nexus: both can run, the plan warns.chaos-meshoverlapslitmus: both can run, the plan warns.cluster-autoscaleroverlapskarpenter: both can run, the plan warns.envoy-gatewayoverlapsingress-nginx,istio-gateway: both can run, the plan warns.falcooverlapsneuvector: both can run, the plan warns.harboroverlapsartifactory,nexus: both can run, the plan warns.ingress-nginxoverlapsenvoy-gateway: both can run, the plan warns.istio-gatewayoverlapsenvoy-gateway: both can run, the plan warns.karpenteroverlapscluster-autoscaler: both can run, the plan warns.kubecost-cost-analyzerconflicts withopencost: the later one is skipped unless--force.kubeflow-pipelinesoverlapsminio: both can run, the plan warns.kubescape-operatoroverlapstrivy-operator: both can run, the plan warns.litmusoverlapschaos-mesh: both can run, the plan warns.neuvectoroverlapsfalco: both can run, the plan warns.nexusoverlapsharbor,artifactory: both can run, the plan warns.opencostconflicts withkubecost-cost-analyzer: the later one is skipped unless--force.trivy-operatoroverlapsneuvector: both can run, the plan warns.vaultoverlapsexternal-secrets,sealed-secrets: both can run, the plan warns.