Skip to content

Data Stream naming is far too restrictive #744

Open
@morrone

Description

@morrone

The approach to Data Stream naming in the connector is too restrictive. Right now the only option we have is to pattern the data stream name like so:

<data.stream.type>-<data.stream.dataset>-<topic>

This pattern doesn't even really allow us to match the common data steam naming scheme which, in kafka terms, would look a bit more like:

<type>-<topic>-<namespace>

And further, the data.stream.type can only be "metrics" or "logs".

None of these restrictions exist in elasticsearch itself; they are only imposed by the connector.

We should really be able to specify the entire data stream name, using some basic substitution variables (for topic name, etc.) in the string.

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