Fully managed (serverless ) ETL service running on Apache Spark environment useful for preparing data for analytics.

AWS Glue discovers data and stores the associated metadata like table definition and schema in AWS Glue Data Catalog (a central metadata repository) by using a crawler.

The crawler will parse data from your connected data sources - data lakes ( S3 ), data warehouses ( Redshift ) or other datastores (RDS, DynamoDB, EC2) - and create Catalog Tables with the representation of the schema and the metadata for that data metadata and information about where source data is stored.

Besides Data Catalog, there are other services within the offering of Amazon Glue, like Studio (that allows you to create, submit and monitor ETL jobs ), Databrew (no-code service for transforming data) , Workflows (to orchestrate ETLs)