Observability Stack & OpenTelemetry Collector

⭐ A local observability backend suite powered by Docker Compose, OpenTelemetry Collector, Prometheus, Grafana Tempo, Grafana Loki, and Grafana.

This repository provides a unified telemetry pipeline designed to collect, process, and visualize Metrics, Traces, and Logs emitted by the Authentication User Dashboard Application.

---

Table of Contents

• Live System Dashboard
• Architecture
• Services Included
• Getting Started
• Ports & Endpoints
• Configuration Details
• License

---

Live System Dashboard

Key Performance & Security Highlights:

• RED Metrics & Traffic: Real-time visibility into HTTP throughput by route and percentile request latency (p50, p95, p99).
• Security Audit Telemetry: Dedicated tracking of auth outcomes, including login successes (200 OK), invalid credentials (401 Unauthorized), CSRF violations (403 Forbidden), and rate-limiting blocks (429 Too Many Requests).
• Correlated Telemetry: Native OpenTelemetry ingestion allowing trace-to-log correlation across backend services.

---

Architecture

flowchart LR
    App["Node.js Application\n(OTEL SDK)"]

    subgraph Collector["OpenTelemetry Collector"]
        Receiver["OTLP Receiver\n(HTTP 4318 / gRPC 4317)"]
        Proc["Memory Limiter & Batch Processors"]
        Receiver --> Proc
    end

    subgraph Storage["Telemetry Storage"]
        Prom["Prometheus\n(Metrics)"]
        Tempo["Grafana Tempo\n(Traces)"]
        Loki["Grafana Loki\n(Logs)"]
    end

    Grafana["Grafana UI\n(Port 3001)"]

    App -->|OTLP Signals| Receiver
    Proc -->|Prometheus Exporter :8889| Prom
    Proc -->|OTLP / HTTP| Tempo
    Proc -->|OTLP / HTTP| Loki

    Prom --> Grafana
    Tempo --> Grafana
    Loki --> Grafana

Loading

Services Included

Service	Image	Container Port	Host Port	Purpose
OTEL Collector	otel/opentelemetry-collector-contrib	4317, 4318, 8889	4317, 4318, 8889	Ingests OTLP signals, processes them, and routes to backends.
Prometheus	prom/prometheus:latest	9090	9090	Scrapes metrics from the Collector's Prometheus exporter.
Tempo	grafana/tempo:latest	3200, 4317, 4318	3300, 4319, 4320	Distributed tracing backend storing trace blocks locally.
Loki	grafana/loki:latest	3100	3200	Log aggregation system for OTLP application logs.
Grafana	grafana/grafana:latest	3000	3001	Visualization UI pre-configured with Prometheus, Tempo, & Loki.

---

Getting Started

Prerequisites

• Docker and Docker Compose installed.
• An application exporting telemetry via OTLP (e.g., OpenTelemetry Node.js SDK).

Usage

1. Clone the repository:

   git clone https://github.com/Mel000000/observability-platform.git
   cd observability-platform

2. Create the grafana-storage volume:

   docker volume create grafana-storage

3. Start the stack:

   docker compose up -d

4. Verify running containers:

   docker compose ps

5. Access the UIs:
   • Grafana: http://localhost:3001
   • Prometheus: http://localhost:9090
   • Collector Metrics Endpoint: http://localhost:8889/metrics

---

Ports & Endpoints

Protocol / Target	Ingestion Port	Usage
OTLP / HTTP	4318	Send traces, metrics, and logs over HTTP from your application.
OTLP / gRPC	4317	Send traces, metrics, and logs over gRPC from your application.
Prometheus Scrape	8889	Scraped every 15s by Prometheus.

---

Configuration Details

1. OpenTelemetry Collector (otel-collector/config.yaml)

• Receivers: Ingests OTLP via gRPC (0.0.0.0:4317) and HTTP (0.0.0.0:4318).
• Processors:
  • memory_limiter: Prevents OOM by setting check intervals and memory limits (256 MiB).
  • batch: Batches telemetry data before sending to exporters for performance optimization.
• Exporters: Routes metrics to Prometheus (:8889), logs to Loki (http://loki:3100/otlp), and traces to Tempo (http://tempo:4318).

2. Prometheus (prometheus/prometheus.yml)

Scrapes metrics exposed by the OTEL collector at otel-collector:8889 every 15 seconds.

3. Grafana Auto-Provisioning (grafana/provisioning)

Data sources are automatically provisioned on launch:

• Prometheus (Default metrics provider)
• Tempo (Trace search and visualization)
• Loki (Log ingestion and querying)

---

License

This project is licensed under the MIT License.