# Observability Platform

Client / Employer: PT Eigen Tri Mathema | Category: Observability | Year: 2025

Unified observability platform spanning metrics, logs, traces, and database performance across multiple Kubernetes clusters.

Tags: Prometheus, Grafana, Loki, Tempo, Thanos, PMM

## Overview

A unified observability platform at Eigen spanning metrics, logs, traces, and database performance across multiple Kubernetes clusters — giving teams a single pane of glass for the entire environment.

## The Problem

Eigen runs a large, multi-namespace Kubernetes environment across multiple clusters. Metrics, logs, traces, and database performance were scattered across separate tools, making it hard to correlate an issue from a slow query to a failing pod to a spike in latency.

## The Solution

A unified observability platform was built by combining Prometheus and Thanos for metrics with long-term storage, Grafana Loki for centralized log aggregation, OpenTelemetry Tempo for distributed tracing, and Percona PMM for database performance monitoring — all surfaced through Grafana dashboards with alerting and notifications.

## Key Outcomes

- **Single Pane of Glass:** Metrics, logs, traces, and DB performance correlated in one place.
- **Cluster-Wide:** Observability across nodes, pods, and namespaces in multiple clusters.
- **Long-Term Storage:** Thanos provides durable, queryable historical metrics.
- **Root Cause:** Trace and log correlation speeds up root-cause identification.

## Project Summary

Eigen's Observability Platform unifies Prometheus/Thanos metrics, Loki logs, Tempo traces, and PMM database monitoring behind Grafana for full-stack visibility.

## Key Features

- Prometheus Metrics
- Grafana Dashboards
- Thanos Long-Term Storage
- Grafana Loki Log Aggregation
- OpenTelemetry Tempo Tracing
- Percona PMM Database Monitoring
- Cluster-Wide Visibility
- Alerting & Notifications
- Historical Querying
- Root Cause Identification