Skip to content

DMS: Improve performance by synthesizing batches #121

@amotl

Description

@amotl

Problem

CTK is currently processing DMS events sequentially, see io/kinesis/relay.py. The consequence is data loading will be slow.

Proposal

Because the ingress DMS event stream can include arbitarily-alternating event types, the data loading in CTK needs a component which synthesizes per-table batches of the same kind of DML operation before submitting them to the database.

It would be nice if this package could provide such a component, because it includes and implements relevant domain knowledge about DMS events already.

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Type

    No type

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions