Skip to content

Lab Series Introduction

Difficulty: Advanced Last Updated: February 23, 2026

Overview

This lab series provides a comprehensive, hands-on journey through building a full-stack observability platform for Kubernetes-based microservices. You will deploy and integrate multiple observability tools across two EKS clusters, implementing the three pillars of observability (Metrics, Logs, Traces) with real-world patterns.

The architecture simulates a production-grade environment with a Managed Cluster hosting the observability stack and a Service Cluster running MSA applications with OTel instrumentation.

Architecture Overview

Architecture Diagram

Prerequisites

Before starting this lab series, ensure you have the following:

RequirementVersionVerification Command
AWS Account-aws sts get-caller-identity
AWS CLI>= 2.15aws --version
eksctl>= 0.175eksctl version
kubectl>= 1.29kubectl version --client
Helm>= 3.14helm version
Terraform>= 1.7terraform version
k6>= 0.49k6 version
Docker>= 24.0docker --version

Required IAM Permissions

Your AWS user/role needs the following permissions:

  • EKS full access
  • EC2 full access (for node groups)
  • VPC full access
  • IAM limited access (for IRSA)
  • CloudFormation full access
  • SQS/SNS full access
  • RDS full access (for Aurora)
  • OpenSearch full access
  • Managed Prometheus/Grafana full access
  • MWAA full access

Cost Estimate

Warning: This lab series creates significant AWS resources. Estimated costs are provided below.

ServiceConfigurationHourly Cost (USD)
EKS Control Plane2 clusters$0.20
EC2 (Managed Cluster)3x m5.xlarge$0.58
EC2 (Service Cluster)3x m5.large (+ Karpenter scaling)$0.29+
Aurora PostgreSQLdb.r6g.large (multi-AZ)$0.52
OpenSearchm6g.large.search (2 nodes)$0.25
Amazon Managed PrometheusBased on ingestion~$0.10
Amazon Managed Grafana1 workspace$0.15
MWAAmw1.small$0.31
SQS/SNSBased on usage~$0.01
Total Estimate~$2.50/hour

Tip: Complete the lab in a single session and run cleanup immediately to minimize costs.

Lab Sequence

PartTitleDurationKey Topics
1Infrastructure Setup60 minEKS clusters, AWS services, ArgoCD
2Observability Stack90 minOTel, Prometheus, Loki, Tempo, Grafana
3MSA Deployment & Canary60 minArgoCD, Argo Rollouts, OTel instrumentation
4Load Testing & Scaling45 mink6, KEDA, Karpenter
5Alerting & AIOps60 minAlertmanager, OnCall, CloudWatch Investigations
6Distributed Tracing45 minTempo, TraceQL, Log-Trace correlation

MSA Application Overview

The lab uses a sample e-commerce MSA application with 5 services:

ServiceLanguageRoleDependencies
API GatewayGoRequest routing, authenticationOrder, Payment
Order ServicePython (FastAPI)Order management, inventoryAurora, SQS
Payment ServiceJava (Spring Boot)Payment processingAurora
Notification ServiceNode.js (Express)Email/SMS notificationsSQS consumer
Analytics BatchPythonDaily analytics aggregationAurora, triggered by MWAA

Service Call Flow

Observability Tool Coverage

This lab covers the following observability tools:

CategoryTools CoveredAWS Integration
MetricsPrometheus, VictoriaMetrics, MimirAMP (remote write)
LoggingLoki, ClickHouse, Fluent BitCloudWatch Logs, OpenSearch
TracingTempo, OTel CollectorX-Ray (via OTel)
VisualizationGrafanaAMG
AlertingAlertmanager, Grafana OnCallCloudWatch Alarms, SNS
AIOpsCloudWatch InvestigationsBedrock Claude integration

Note: This lab focuses on open-source and AWS-native tools. Commercial solutions like Datadog and Dynatrace are covered in separate documentation but not deployed in this lab.

Learning Outcomes

By completing this lab series, you will be able to:

  1. Design a production-grade observability architecture for Kubernetes
  2. Deploy the complete LGTM stack (Loki, Grafana, Tempo, Mimir) with OTel
  3. Configure multi-backend telemetry pipelines using OTel Collector
  4. Implement canary deployments with observability-driven analysis
  5. Build AIOps workflows with CloudWatch Investigations and Bedrock
  6. Analyze distributed traces to identify performance bottlenecks
  7. Correlate metrics, logs, and traces for root cause analysis

References


Ready to begin? Start with Part 1: Infrastructure Setup