DuckLake Container - Data Lakehouse on S3

THE PROBLEM Setting up a data lakehouse shouldn't take a week. But if you've tried to configure Apache Iceberg with Hive Metastore, or wrestled with AWS Glue catalogs, you know the reality: days of YAML files, IAM policies, Spark clusters, and debugging before you can write your first query. For large enterprises with dedicated platform teams, that's fine. For a 5-person startup or a solo data engineer, it's a blocker. THE SOLUTION DuckLake Container gives you a fully functional data lakehouse in one container. PostgreSQL stores your catalog metadata. S3 stores your data as Parquet files. You connect from any DuckDB client and start querying. Time to first query: 5 minutes. No Spark. No Hive. No Glue. No YAML sprawl. HOW IT WORKS

Subscribe and launch the container on ECS, EKS, or Fargate Point it at your S3 bucket via environment variables or Secrets Manager Attach an IAM role with S3 permissions Connect from any DuckDB client and start querying

Your data lives in S3 as standard Parquet files. The catalog lives in PostgreSQL. Both stay in your VPC we never see your data. WHAT YOU GET Lakehouse capabilities: ACID transactions on S3, time travel queries, schema evolution, and partition pruning for fast queries on large datasets. Operational simplicity: Single container deployment with no external dependencies beyond S3. Automatic IAM credential handling no access keys to manage. Configuration via environment variables or AWS Secrets Manager. Flexible connectivity: Connect from Python, Go, Rust, Java, or any language with DuckDB bindings. Works with DataGrip, DBeaver, or any SQL tool. Multiple concurrent clients supported. WHAT'S INSIDE

PostgreSQL 17 for catalog storage (table schemas, snapshots, partitions) DuckDB 1.3 with ducklake, postgres, and httpfs extensions pre-configured Python init script that validates S3 access and configures the catalog Built-in container health monitoring

WHO THIS IS FOR Small data teams (1-10 people) who need lakehouse capabilities but don't have time to configure and maintain Iceberg, Delta Lake, or Glue. Startups who want to start simple and scale later. Your data is in standard Parquet format if you outgrow this solution, migrate to Iceberg or Databricks without rewriting everything. Developers building analytics features who need a quick way to prototype lakehouse functionality. Data engineers who are tired of complexity. If you've ever spent a week debugging Hive Metastore or Glue permissions, this is for you. WHO THIS IS NOT FOR Large enterprises already invested in Databricks, Snowflake, or a mature Iceberg deployment. This is a lightweight solution it won't replace your existing data platform. High-concurrency production workloads with hundreds of concurrent writers. PostgreSQL catalog handles moderate concurrency, but this isn't designed for massive scale. SECURITY Runs in your VPC your data never leaves your AWS account. Uses IAM-based authentication with task roles, no hardcoded credentials. Works with private S3 buckets. Supports AWS Secrets Manager for secure configuration. PRICING $0.05 per hour, billed per second with a 1-minute minimum. Running 4 hours/day costs approximately $6/month. Running 8 hours/day costs approximately $12/month. Running 24/7 costs approximately $36/month. Plus standard AWS costs for ECS/EKS compute, S3 storage, and data transfer. No annual commitment. No minimum spend. Stop the container, stop paying. GETTING STARTED

Subscribe to this listing Create an S3 bucket for your data Create an IAM role with S3 read/write permissions Launch the container with your configuration Connect and query

Documentation: https://docs.lokryn.com/ducklake  ABOUT LOKRYN Lokryn builds infrastructure tools for small dev teams the kind of stuff that usually costs $10K/month from enterprise vendors or takes weeks to configure from open source. Simple. Transparent pricing. Runs in your cloud.