sayyed-mujjakir/streamflow — reverse-engineered prompt

Build me a production style real time event pipeline on AWS called StreamFlow.

I want a small Python producer that sends fake events into Kinesis, and a worker service that reads them, enriches or processes them, then stores results in Postgres, Redis, and S3. The worker should expose Prometheus metrics so I can see processed event counts, latency, and errors.

Set up all the cloud infrastructure with Terraform, including VPC, EKS, Kinesis, RDS Postgres, ElastiCache Redis, S3 buckets, IAM roles for pods, Secrets Manager, and separate dev and prod config. Package the worker for Kubernetes with Helm, include autoscaling, and make sure it can run safely without hardcoded AWS keys.

Also add monitoring config for Prometheus and Grafana style alerts, plus a GitHub Actions pipeline that tests, validates Terraform, builds Docker images, runs plan and apply with approval, and deploys to EKS.

Please include clear README steps for local testing with LocalStack and for deploying to AWS.