Aws Observability Labs

AWS Observability Labs


Six Terraform-driven AWS CloudWatch labs covering governance, dashboards & alarms, CloudWatch Agent, EMF/PutMetricData custom metrics on a Flask Orders API, dynamic Auto Scaling (target / step / simple), and Bedrock Claude AI log analysis rendered inline on a dashboard.

---

Purpose

A progressive series that tells one coherent monitoring story: how to go from a bare AWS account with no visibility to a production-grade observability posture with AI-assisted log analysis — every step in Terraform, every step deployable and destroyable in isolation.

The labs build on each other conceptually (not operationally — each lab is self-contained and owns its own VPC, dashboard, and alarms) so you can deploy a single lab to learn a specific technique, or walk the full arc:

**Govern → Observe → Instrument → Measure → Scale → Analyze**

The finance-domain framing is deliberate: Labs 02 / 04 / 05 / 06 wire a trading-system workload (DynamoDB `trade-executions`, Flask Orders API, financial-orders SQS queue, trade-processor Lambda) into the CloudWatch story so the demos are closer to the problems a real data-engineering or cloud-architecture client would bring.

---

Labs

#	Lab	Description	Key Services
01	Governance Monitoring Overview	ASG + CPU alarms + SNS → Lambda (CloudTrail-enriched notifications) + CloudTrail trail	EC2, ASG, CloudWatch Alarms, SNS, Lambda, CloudTrail
02	CloudWatch Metrics, Dashboards & Alarms	Multi-service dashboard over EC2 / DynamoDB / S3 with five alarm types (threshold, stop-action, anomaly detection, DynamoDB spike, S3 delete audit) plus a composite alarm	CloudWatch, EC2, DynamoDB, S3, SNS
03	[EC2 Detailed Monitoring + CloudWatch Agent](lab