A development and learning observability platform built on OpenTelemetry, Grafana, Loki, Prometheus, and Tempo. Perfect for local labs, learning, and proof-of-concepts. Deployed via ArgoCD for GitOps management with local filesystem storage and multi-tenant support.
⚠️ Not for Production: This setup is designed for development, learning, and lab environments. For production deployments, additional security, scaling, and operational considerations are required.
- OpenTelemetry Collector - Unified telemetry data collection with intelligent routing
- Grafana - Pre-configured dashboards with multi-tenant data sources
- Loki - Scalable log aggregation with multi-tenant support and optimized resource usage
- Tempo - High-scale distributed tracing with local filesystem storage
- Prometheus - Metrics collection with remote write support
- ArgoCD - GitOps deployment and management
- Multi-Tenant Architecture - Separate data isolation with 'foo' and 'bazz' tenants for logs
- Resource Optimized - Runs efficiently on local development machines
- Lab-Friendly - Easy setup for development, learning, and testing
- Educational - Great for understanding observability and multi-tenancy concepts
# Clone and install (read-only setup)
git clone https://github.com/fiddeb/observabilitystack.git
cd observabilitystack
./scripts/install_argo.shPerfect for learning and experimentation:
# 1. Fork this repository on GitHub (click Fork button)
# 2. Clone your fork (replace YOUR_USERNAME)
git clone https://github.com/YOUR_USERNAME/observabilitystack.git
cd observabilitystack
# 3. Setup ArgoCD to use your repository
./scripts/setup_argocd.sh
# 4. Install everything
./scripts/install_argo.shWhy fork? Customize configurations, experiment safely, and contribute back via pull requests.
That's it! Your observability stack is now running with:
- Grafana at http://grafana.k8s.test (admin/admin)
- ArgoCD at http://argocd.k8s.test (admin/)
- Prometheus at http://prometheus.k8s.test
- Loki at http://loki.k8s.test
- Tempo at http://tempo.k8s.test
- OpenTelemetry Collector at http://otel-collector.k8s.test
Data Flow:
- Logs → OpenTelemetry Collector → Loki → Local Storage → Grafana
- Metrics → OpenTelemetry Collector → Prometheus → Local Storage → Grafana
- Traces → OpenTelemetry Collector → Tempo → Local Storage → Grafana
Verify everything works end-to-end:
# Run automated tests
kubectl apply -f manifests/telemetry-test-jobs.yaml
# Check results in Grafana
"Metrics: Navigate to Prometheus → gen{}"
"Logs: Navigate to Loki → {job=\"telemetrygen-logs\"}"
"Traces: Navigate to Tempo → {service.name=\"telemetrygen\"}"observabilitystack/
├── argocd/ # ArgoCD application definitions
│ └── observability-stack.yaml
├── docs/ # Documentation
│ ├── ARCHITECTURE.md # Design & concepts (umbrella chart pattern)
│ ├── INSTALLATION.md
│ ├── USAGE_GUIDE.md
│ └── ...
├── helm/stackcharts/ # Helm umbrella chart
│ ├── Chart.yaml # Chart dependencies
│ ├── values/ # Split configuration (one file per component)
│ │ ├── base.yaml # Component enable/disable flags
│ │ ├── loki.yaml # Loki configuration
│ │ ├── tempo.yaml # Tempo configuration
│ │ ├── prometheus.yaml # Prometheus configuration
│ │ ├── grafana.yaml # Grafana configuration
│ │ ├── mimir.yaml # Mimir configuration
│ │ ├── garage.yaml # Garage S3 storage profile (opt-in)
│ │ ├── perses.yaml # Perses configuration
│ │ ├── spaceport.yaml # Spaceport configuration
│ │ └── opentelemetry-collector.yaml # OTel configuration
│ └── charts/ # Downloaded dependency charts (.tgz)
├── manifests/ # Kubernetes manifests
│ ├── argocd-ingress.yaml # ArgoCD web access
│ └── telemetry-test-jobs.yaml # Test workloads
└── scripts/ # Automation scripts
├── install_argo.sh # Complete installation
├── force_argo_sync.sh # ArgoCD sync management
├── setup_garage.sh # Initialize Garage object storage
├── merge_feature.sh # Git workflow
└── test_multi_values.sh # Validate configuration
The stack uses a multi-values approach for better organization:
helm/stackcharts/values/base.yaml- Control which components are installed- Component-specific files - One file per component (loki.yaml, grafana.yaml, etc.)
Example: Disable Tempo to save resources
# Edit helm/stackcharts/values/base.yaml
tempo:
enabled: falseSee Architecture Guide for detailed configuration patterns.
- Architecture Guide - Umbrella chart pattern, multi-values configuration, and customization
- Installation Guide - Complete setup instructions
- Usage Guide - How to send and query telemetry data
- Git Workflow - Branch management and deployment
- Troubleshooting Guide - Emergency procedures, debugging, and complete command reference
- force_argo_sync.sh - Intelligent ArgoCD sync
- merge_feature.sh - Safe feature branch merging
# Health check and sync
./scripts/force_argo_sync.sh
# Get ArgoCD admin password
kubectl -n argocd get secret argocd-initial-admin-secret -o jsonpath="{.data.password}" | base64 -d && echoQuick fixes:
# Not working? Run the health check
./scripts/force_argo_sync.sh
# Still issues? Use port forwarding
kubectl port-forward service/grafana 3000:80 -n observability-lab &
kubectl port-forward svc/argocd-server -n argocd 8080:443 &Need help? Check Troubleshooting Guide for emergency procedures and debugging commands.
We welcome contributions! See our Git Workflow for development practices.
# Create feature branch
git checkout -b feat/my-awesome-feature
# Make changes and test
./scripts/force_argo_sync.sh
# Merge safely
./scripts/merge_feature.sh feat/my-awesome-feature
or create pull request :)This project is licensed under the MIT License. See the LICENSE file for details.