This repository contains the implementation of a monitoring system for microservices running on a Kubernetes cluster. The project integrates various tools and practices for infrastructure management, application deployment, and system monitoring.
- Objective: Build a monitoring and tracing system for microservices running on a Kubernetes cluster with 4 nodes (1 master and 3 worker nodes).
- Scope:
- Monitor performance metrics of Kubernetes and microservices.
- Centralize logging and enable tracing for quick troubleshooting.
- Deploy a scalable and resilient monitoring stack.
-
Infrastructure as Code:
- Utilized Terraform to provision infrastructure on Google Cloud Platform (GCP).
- Configured network, firewalls, and Load Balancer for the Kubernetes cluster.
-
Kubernetes Cluster:
- Set up a Kubernetes cluster from scratch using Ansible.
- Deployed microservices (e.g.,
auth-svc,gateway-svc,user-svc) as Helm charts.
-
Monitoring Stack:
- Tools used: Prometheus, Grafana, Tempo, Loki, and Promtail.
- Configured:
- Prometheus for collecting metrics.
- Grafana for visualization dashboards.
- Loki for centralized logging.
- Tempo for tracing.
- Integrated with OpenTelemetry Java Agent for tracing in Spring applications.
-
Visualization and Alerts:
- Built Grafana dashboards for real-time monitoring.
- Set up alerts to notify the team about critical issues.
- Languages: Java (Spring Framework), YAML (Helm charts)
- DevOps Tools: Terraform, Ansible, Helm
- Monitoring Tools: Prometheus, Grafana, Loki, Tempo, Promtail
- Cloud Provider: Google Cloud Platform
- Overcame the complexity of setting up a custom Kubernetes cluster without using managed services like GKE.
- Learned to integrate and optimize monitoring tools for a production-like environment.
Feel free to explore the repository and provide feedback!