● CLUSTER OPERATIONALLIVE · SyncedK3s v1.34.4+k3s1 HA

Mateere Cloud Lab

Public curated infrastructure showcase for Mateere Cloud Lab. Built and operated by Frank Chinembiri. Demonstrates multi-node hybrid Kubernetes orchestration, host-level CloudPanel SNI coexistence, automated GitOps delivery, and everyday service reliability.

Executive Summary

Mateere Cloud Lab began as a personal passion project to explore systems engineering from the ground up. Today, it hosts applied Earth Observation research (GeoCrop), development compute for platforms I lead (Lead Engineer & DevOps Engineer on the Next-Gen school management platform), and daily family services. It reflects hands-on competence across Linux networking (Flannel CNI, UFW firewall sync), zero-downtime certificate automation (cert-manager), declarative GitOps pipelines (ArgoCD & Gitea), and air-gapped security models.

Read the 11-Step Origin Story Interactive Architecture Service Fleet (141)

Curated Infrastructure Telemetry

Source of Truth: snapshot.json · Verified Real Data

Nodes

3 Control + 1 Worker

Src: k3s_api_server

Workloads

185

Across 19 namespaces

Src: kube_state_metrics

Services

141

Cataloged endpoints

Src: service_discovery

Endpoints

Synthetic probes

Src: uptime_kuma_api

30d Availability

74.47%

Cluster-wide aggregate

Src: uptime_kuma_30d

Cluster Uptime

189d

Zero node reboot

Src: k3s_server_uptime

Platform Health: 97.16% Across 45 Monitored Synthetic Endpoints

Production microservices maintain 99.8%+ uptime; 97.16% reflects planned overnight maintenance on research workers and experimental nodes.

PROVENANCE: Uptime Kuma 30d Window

0%97.16% Live Operational Fleet100%

Multi-Node Hybrid Cluster Topology

Physical edge routing, host-level Nginx SNI demuxing, and distributed K3s node architecture.

K3s HA Consensus + Calico CNI

Cluster Capacity Utilization

Telemetry Snapshot

Cluster CPU Utilization30% (22 Cores)

Cluster Memory Utilization70.6% (42 GB)

Active Storage Pools0.91 TB Local NVMe/SSD

Memory overhead accommodates persistent Redis queues, PostGIS GIS indices, and in-memory spatial raster caching.

Workloads by Namespace (185 Total)

19 Namespaces

nextgen

family-apps

geocrop

monitoring

cert-lab

kube-system

supabase

authentik

other (11 namespaces)

Strict cgroup resource quotas (`limits`/`requests`) prevent noisy neighbor cross-contamination.

Flagship Workloads & Realtime Health

Live service status across the 6 infrastructure tiers.

View Full 141 Services Directory →

CertLab Simulator & AcademyUP

FastAPI · 12 Certs · 47 Guided Chapters · 80% Mastery

ns: cert-lab100% (30d)

CK-X Kubernetes Lab SimulatorUP

ttyd · 19 CKA/CKAD Mock Exams · Auto-Grading

ns: cert-lab100% (30d)

GeoCrop Inference WorkerUP

FastAPI · Redis RQ · MinIO

ns: geocrop99.8% (30d)

Next-Gen Platform & APIUP

School Mgmt · Lead & DevOps · Supabase

ns: nextgen99.9% (30d)

Stagri Precision PortalUP

CloudPanel · Nginx · PHP

ns: host99.7% (30d)

ArgoCD GitOps SyncUP

ArgoCD · Kubernetes Kustomize

ns: argocd100% (30d)

Gitea Source ControlUP

Gitea · SQLite · OCI Container

ns: geocrop99.9% (30d)

Authentik Identity SSOUP

Forward-Auth · OIDC / SAML

ns: authentik99.9% (30d)

Prometheus TSDB FleetUP

Prometheus · cAdvisor · Pushgateway

ns: monitoring99.9% (30d)

Uptime Kuma SyntheticUP

Uptime Kuma · 45 Probes

ns: monitoring99.9% (30d)

MinIO S3 Object StorageUP

S3 API · COG GeoTIFF Store

ns: geocrop100% (30d)

PostgreSQL StatefulSetUP

PostGIS · WAL Replication

ns: supabase99.9% (30d)

Workload Tiering & Classification Model

Transparently categorizing commercial, research, platform, and domestic infrastructure.

[ PROFESSIONAL ]

Active commercial and client production systems operated under SLA commitments.

Examples: Next-Gen School Management Platform, Stagri, Green Earth backend services

[ RESEARCH ]

Applied machine learning and Earth Observation research inference pipelines.

Examples: GeoCrop Inference Worker, Sen4Stat EO Engine, DEA STAC ingestion

[ PLATFORM ]

Core cluster delivery, identity management, ingress controllers, and consensus engines.

Examples: ArgoCD, Gitea SCM, Authentik SSO, MinIO Object Storage, PostgreSQL

[ OBSERVABILITY ]

Telemetry collection, metric timeseries databases, and synthetic uptime probes.

Examples: Prometheus Server, cAdvisor, Node Exporter, Uptime Kuma, Umami

[ FAMILY ]

Real-world household services operated continuously for actual everyday users.

Examples: Nextcloud Storage, Jellyfin Streaming, Mealie, Grocy, FarmOS

[ EXPERIMENTAL ]

Temporary sandbox spikes, transient branch previews, and ephemeral testing containers.

Examples: Ephemeral feature branches, prototype microservices

Domestic & Farm Operations Utility

Infrastructure With Real Consequences

Operating real services for real people: Nextcloud for family records, Jellyfin for local media streaming, Mealie & Grocy for household inventory, and FarmOS for agricultural crop tracking. If the cluster drops, family life is interrupted.

NextcloudJellyfinGrocyMealieFarmOS

The Personal Journey

From One Server to a 4-Node Cloud Platform

Discover the complete 11-step evolution: beginning with enterprise HPC operations at ZCHPC, confronting single-node memory and port bottlenecks, adopting K3s orchestration, and scaling into the multi-node platform running today.

Read Full 11-Step Story